Answering Machines and Voice Mail


Problem: Answering machine all the time

Question:
Sometimes we make many calls to a number, and we always reach an answering machine. What should we do?

Answer:
If the answering machine message indicates that the selected number is a household, we either must get through or else must count the case as an indirect refusal. So it is worth making many calls to complete the case, since we already know that the phone number is for a household.
From a practical point of view, however, we should not make an excessive number of calls if we have limited resources (which is always the case). Past experience with answering machines suggests the following:
1) Most people who use answering machines are willing to do an interview if we can reach them. And a good time to reach people who use answering machines is Saturday morning. Consequently, we should try two Saturday morning calls for phone numbers with answering machines.
2) After reaching an answering machine 10 times, the chances of completing an interview are very small. At that point, if we have made two Saturday morning calls, we should give up on the case and count it as an indirect refusal.
(TP 2/22/96)


Problem: Answering machine for a business

Question:
Calls to this number get the message, "Nanette Design Group; I'm either on the other line or out of the studio." Could this be a residence?

Answer:
It is possible that this is a number for a business operated out of a person's home and that the phone number is also for personal use. We should try to get through, in order to determine if anyone lives at the location where the phone rings. Specifically, we should make at least two calls during normal business hours, plus two calls during the evening, and a call on the weekend. If no one ever answers and we still keep getting the same message, we can finalize the case as Not-a-residence, provided that the content of the message indicates that it is a business.
(TP 2/15/96)


Problem: Answering machine refers to another number

Question:
We reached an answering machine that said to "call me at xxx-xxxx." Should we call that other number and try to do an interview there?

Answer:
Yes, call the number given in the message and try to clarify the situation.
If the original sampled number was a residence, we would attempt to enumerate all eligibles who reside in that household. We would not include anyone at the new number who did not reside in the original sampled household.
If, on the other hand, the original sampled number was NOT a residence, we would NOT attempt an interview, even if the new number was a residence. The reason is that the new number has its own chance of being selected, and we do not want to change its probability of selection because of this answering machine message.
(TP 2/6/96)


Problem: Answering machine confirms different number

Question:
We reached an answering machine, and the message says we have reached a number that is different from the sampled number that we called. Should we continue calling this number?

Answer:
Call the number mentioned in the answering machine message, and try to clarify the situation.
If you get the same message at the new number, it indicates that the sampled number, and perhaps other numbers as well, are part of a seek group used by a business or government office; or this could be a case of crossed lines. In either case the sampled number can be coded Not-a-residence.
If, on the other hand, you do NOT get the same message at the new number, or the new number is not a working number, the answering machine message may be an attempt to discourage calls. Continue calling the sampled number as you would any phone number with an answering machine.
(TP 2/6/96)


Problem: Answering machine then disconnected

Question:
An answering machine was reached 7 times at the sampled number, and on the 8th call the number was disconnected. The answering machine message sounded like it was for a residence. No enumeration was ever made. Should we code this a refusal or a Not-in-service?

Answer:
Code this an enumeration refusal. It was most likely an eligible household that we simply were not able to get through to.
(TP 2/5/96)


Problem: Answering machine for a switchboard

Question:
At one number we reach an answering machine that says, "Thank you for calling Marisol; please enter the extension of the party you would like to speak with now." You can't leave a message unless you know an extension number. This is almost certainly a business; can we finalize it as such?

Answer:
Yes, code this as Not-a-residence. It is not necessary to make repeated calls to this number.
(TP 1/22/96)


Problem: Answering machines with no message

Question:
We often have phone numbers that we call many times and that ring with no answer. But then on a subsequent call we'll reach strange noises which the interviewer can't identify. If we call the operator, we are often told that these are "messageless answering machines." I've seen this enough to have reached the conclusion that operators just say this when they don't know what else to say. It seems to me that these aren't valid numbers.

Answer:
Some non-working numbers produce funny signals that can be confused with answering machine beeps. Unless there is a message, however, it is not likely to be an answering machine. If no one has ever answered such a number, it can be coded Not-in-service after 3 outcomes that result in those funny signals. Or if we have made 18 calls that have resulted in a combination of no-answers and a few funny signals, the phone number can be coded as Assumed-not-a-residence.
(TP 2/13/96)


Problem: Answering machine for a fax

Question:
We reach an answering machine which says, "Hello this line is being answered by a facsimile machine. If you wish to fax, you can begin your transmission now. If you do not wish to fax, please check the number and try your call again." What do you suggest?

Answer:
Although fax machines are sometimes found in residences, this message indicates that the phone number is not used at all for voice transmission. If no one has ever answered this number, code it as Not-a-residence after getting that message 3 times.
(TP 2/16/96)


Problem: Voice mailbox services

Question:
At some numbers we get messages such as, "This is the Communications Gateway; please enter your mailbox number," or "please enter the code for the party you are trying to reach." Are these residential numbers?

Answer:
These are message services. Code such phone numbers as Not-a-residence. The individuals who use them for personal calls are likely to have their own (possibly unlisted) residential numbers as well.
(TP 1/22/96)


Problem: Voice mail discontinued

Question:
We reach a recording which says, "Voice mail service for this telephone number is temporarily not available; consult your directory for the business you would like to reach." How should we code this?

Answer:
This may be the number of a private message service, in which case we would code the number as Not-a-residence. It could also be a disconnected business number, in which case the appropriate code would be Not-in-service. Since telephone companies do not usually provide messages for disconnected business numbers, I would guess that the number belongs to a private message service company. In cases like this, choose one of these two outcome codes, and do not waste time trying to clarify this matter. Neither outcome code will count against the response rate.
(TP 2/6/96)


Problem: Voice mail plus regular line

Question:
We completed an interview with a person who told us that he uses two telephone lines -- one with an unlisted number that he rarely gives out, and another that goes to a voice mail service that he uses to screen his calls. We reached him on his unlisted personal number. As part of the interview we ask how many separate phone lines come into the respondent's household. Should I count his voice mail line, and record that he has 2 lines?

Answer:
Yes, count both lines. Since we would select this household by selecting either number, there are effectively two "pipelines" into the household. Note, however, that this would not apply if the voice mail line belonged to a business outside the respondent's home. Phone numbers from someone's place of work are not counted as household numbers.
(TP 2/6/96)


Cellular and Mobile Phones


Problem: Reached cellular phone

Question:
Someone answered and confirmed the sampled number, and it was a cellular phone. Should we interview that person?

Answer:
Usually not. We would only attempt to interview that person in the unlikely case in which the person who answered the cellular phone has no other residential phone. Ordinarily, however, we would consider a cellular telephone number as Not-a-residence. That person's chance to be selected for the study is associated with his or her residential telephone, and the use of extra cellular telephone numbers would change our sampling fractions in ways that we could not readily adjust for. We can handle multiple phone numbers in a household, because we ask all respondents how many different lines they have at home. It would not be so simple, however, to determine for each person how many possible numbers that person could be reached at, if we included cellular (and work) numbers.
(TP 2/5/96)


Problem: Cellular informant gave home phone

Question:
At the sampled number we reached a woman on a cellular phone. She gave us her home number and told us we could call her there for an interview. Should we call the home number?

Answer:
No, we should not call the home number unless that number was actually selected in the sample. If we used referrals to a home number from a cellular phone, a work phone, or any other source, it would produce unequal selection probabilities that we could not adjust for. Code the cellular phone number as Not-a-residence.
(TP 2/6/96)


Fax and Computer Lines


Problem: Modem signals and busy

Question:
At the sampled phone number we have reached only busy signals and fax or modem signals. How many calls do we need to make to this number?

Answer:
There are two limits that come into consideration here. Fax or modem signals are sometimes encountered at a residential number, but they usually indicate that a number is not a residential one. After getting a modem signal 3 times, code the number Not-a-residence, provided that no one has ever answered and no answering machine with a message was ever encountered.
The other limit to be considered is on the total number of calls. If 18 calls have been made, including busy, modems, no-answers, and any other outcomes, the number can be coded Assumed-not-a-residence, provided again that no one has ever answered and no answering machine with a message was ever encountered.
(TP 2/16/96)


Problem: Apparent fax or modem signals

Question:
We occasionally have a phone number that we have called many times and no one has ever answered, but interviewers have coded the outcomes of two or three of the calls as "modem/fax". I think these are strange phone signals and not really modems. In cases like these, I think it's pretty clear that, whatever is going on, they are not real phone numbers. Do you have any general guidelines on how to handle situations like this?

Answer:
Some non-working numbers produce funny signals that can be confused with fax or modem signals. If the phone number usually rings and no one ever answers, we should go through our usual cycle of 18 calls. We then finalize the case as Assumed-not-a-residence. The presence of a few of these funny signals, mixed in with the no-answers, is not a sufficient reason for us to change that procedure. If anything, they only bolster our conclusion that these are not working residential numbers.
(TP 1/22/96)


Problem: Computerized thank you

Question:
At one number we get a computerized voice that says, "thank you," followed by a quick series of tones. What is this?

Answer:
This is probably a pay phone that does not accept in-coming calls. If no one has ever answered this number, code it as Not-a-residence after getting that message 3 times.
(TP 2/16/96)


Multiple Phones in a Household


Problem: Teenager's phone

Question:
We reached a telephone that was used by teenage children in a household. They are too young to interview, so what should we do?

Answer:
Ask to speak to an adult so that you can enumerate the eligible persons in the household and select a respondent to interview. It is permissible to call the household back on a phone number usually answered by adults, if that will make things easier. The main point to bear in mind is that each household has one or more "pipelines" into it -- the telephone numbers that ring in that household. If any one of those numbers is selected, the entire household is selected. Households with more than one phone line have more chances to be selected, but we adjust for that difference by asking each respondent how many phone lines come into the household, and we use that information to create weights. A telephone used by teenagers is just one more "pipeline" into that household.
(TP 2/5/96)


Problem: Refusal on teenager's phone

Question:
The selected telephone number reached a teenager's phone. His parents live in the same household upstairs. The teenager told us that he had spoken with his mother about the study, and she decided that she did not want to get involved and did not give permission to have the upstairs number given out. Since it is a minor refusing, how should we finalize the case?

Answer:
This is an enumeration refusal, in spite of the young age of the informant. Notice that the teenage phone is a legitimate "pipeline" into this household, and we would have enumerated all eligible adults and selected a respondent if possible.
(TP 2/5/96)


Problem: Telephone in a home office

Question:
An informant told us that we had reached a phone that was located in a residence but was used for a business operated out of the home. Should we consider this a residential phone and attempt to enumerate the household?

Answer:
Yes, a telephone in a home office is considered one more "pipeline" into the household, the same as an extra line for teenagers. As a consequence, we should be sure to include home office lines when we ask respondents how many different telephone lines come into their households.
(TP 2/15/96)


Problem: Line used only for computer or fax

Question:
As part of the interview we ask how many phone lines come into the respondent's household. Some people have a separate line just for their computer or fax machine. Should we include that line in the total for the household?

Answer:
It depends if the phone on that extra line is set up to ring. Remember that what we are trying to do is to establish the number of different "pipelines" into the household.
If the extra line is plugged directly into a modem or fax and never rings, then it is not another possible means of access for our interviewers and should not be counted as an additional household line.
On the other hand, if the extra line is connected to a regular phone and can be answered when it is not being used by the modem or fax, then it should be counted as an extra household line.
(TP 2/16/96)


Problem: Business line forwarded to home

Question:
We reached an informant who told us that the sampled phone number belongs to his business telephone, which is forwarded to his home after he leaves the office. Should we proceed with enumeration of the household?

Answer:
No, we should not attempt an interview if the sampled phone number belongs to a telephone located outside a residence, in spite of the fact that the number sometimes rings in a household. This should be coded Not-a-residence. Note, however, that a telephone in a home office is treated the same as a residential line.
(TP 2/15/96)


Problem: Phone number rings in two locations

Question:
We have been told by an informant that the sampled phone number rings in her home, but it also rings in the home of the people that the sampled number belongs to. (The informant has a different line for her own household.) This informant "always" answers the sampled phone because the owners don't speak English and never answer the phone. In previous calls we have been told that this number is a business number but that people do live there. What should we do?

Answer:
Treat this number as if you had reached a household in which no one speaks English. Do not attempt to interview the informant, since she has a different residential phone number. Households in which no one speaks English are sometimes excluded from the survey population by design, but it depends on the study.
(TP 2/1/96)


Unusual Household Situations


Problem: House-sitters

Question:
An informant told us she was house-sitting, and the residents won't be back for a couple of months. Should we try to interview the house-sitter, or what should we do?

Answer:
The answer depends on whether or not the house-sitter has another residence that can be considered that person's permanent residence.
1) If the house-sitter has another residence, he or she is not enumerated as part of this household; the case can be finalized as out-of-town for the duration, if the residents are not expected to return until after the study is over. If there is some doubt about whether the residents will return before the study ends, try reaching them later.
2) If the house-sitter does NOT have another residence, enumerate that person as part of the household, together with the absent residents, and select one of the enumerated persons as the respondent. If the house-sitter is selected, attempt to interview that person. If one of the absent residents is selected, finalize the case as out-of-town for the duration (or call back later, if there is is a possibility that the absent resident will return before the study ends).
(TP 2/13/96)


Problem: House-sitter then disconnected

Question:
On a previous call we reached a male informant who said that he was house-sitting and that the residents would be back in a couple of weeks. We set a callback for a date after the expected return. But when we called, the number had been disconnected. Should we code this Not-in-service, or a refusal, or what?

Answer:
Code this case as out-of-town for the duration of the study. This should not be coded Not-in-service, since the number was working when we first called. It is not a direct refusal either, although the out-of-town code will have the same effect on the response rate.
(TP 2/6/96)


Problem: Residents away on sabbatical

Question:
After many no-answers, we happened to reach an informant who does security checks on the house. He told us that the residents are away for 6 months. How should we code this?

Answer:
Code this as respondent out of town for duration of the study. Even though we had not enumerated the household and selected a respondent, that is the appropriate code.
(TP 2/6/96)


Problem: House temporarily rented out

Question:
An informant confirmed the sampled phone number and told us that this wasn't her home phone. She said that we had reached a "rental unit," that the owners lived there, but that they had "rented it to several parties." She said we probably wouldn't be able to reach the owners for a few months. What should we do?

Answer:
Since the informant said this was not her home phone, we should not attempt to interview her. However, the situation needs further clarification. If renters are living there who do not have any other residence, we should enumerate them and select one for interviewing. It seems, however, that the rentals are short-term and more like a vacation situation; in such cases we do not attempt to interview renters. The appropriate persons to interview are probably the owners, and they are out of town. If they are away until the study is over, finalize the case as out-of-town for the duration. If there is any doubt about how long they will be out of town, try the number later in the study.
(TP 2/5/96)


Problem: Caretaker for elderly person

Question:
We've spoken to a woman at this number who says she works for a nursing care agency and provides care to the 81-year-old woman who actually lives there. The informant is there during most days of the week, eats there, sleeps there, and goes home for the weekends when a weekend worker takes over. She does not consider this to be her home, even though she spends less time at her "usual home" than she does here. Should this caretaker be included in the enumeration?

Answer:
No, the caretaker should not be included in the enumeration. She has a usual residence elsewhere, and her chance to be selected for the study is associated with her home telephone number. The only eligible respondent at the sampled household is the elderly woman.
(TP 2/5/96)


Problem: Very big family

Question:
We reached a family household with more than 10 adults. Can we just exclude this as a group quarters?

Answer:
If 10 or more UNRELATED adults live in a group setting, we generally call it a group quarters and code the case as Not-a-residence. If, however, some of the adults are related and there are fewer than nine adults who are unrelated to one another, it is just a big family situation, and we should try to enumerate the household and select a respondent.
As a practical matter, our CATI instrumentation only provides for 10 eligibles in the section on household enumeration, so we will have a problem if there are more than 10 eligible persons. But just enumerate the 10 oldest persons, select one of those 10, and write a note on the situation. Households of this size occur so infrequently that the bias associated with missing a few eligibles in such households is not a major problem.
(TP 2/6/96)


Problem: People about to move out

Question:
Female informant answered and confirmed phone number but said they were just about to move out and didn't have time to participate. She said they wouldn't be living there anymore after tomorrow and that some new people would be moving in. Assuming this is true (and not some kind of dodge), can we keep calling and try to enumerate with the new residents, or should we final this as a refusal?

Answer:
This is an enumeration refusal. Besides, if new residents move in, they will most probably have a different phone number.
(TP 8/8/94)


Problem: Phone number disconnected after enumeration

Question:
We enumerated a household at this number and selected a respondent, but we did not do the interview at that time. Now we are getting a message that the number has been disconnected. How do we code this?

Answer:
This should be coded Respondent-never-located, and it has the same effect on the response rate as a refusal. We would have interviewed the selected respondent, but we didn't get to that person in time.
(TP 2/5/96)


Problem: Phone number changed after enumeration

Question:
We enumerated the household at the sampled number. Now we're getting a message saying the phone number has been changed to a new number Are we allowed to call this new number as long as a respondent has been selected?

Answer:
No, we should not call the new number; this case should be finaled as respondent-never-located. We would have interviewed the selected respondent, but we missed our chance. Following respondents when they move is problematic in a sample based on households.
(TP 2/6/96)


Problem: Construction worker at a house

Question:
A male informant was reached at the sampled number. He said that the house was in the process of being reconstructed and that it is presently not lived in. He said that he just happened to be there setting up for the next day's work but that no one would actually be living there for at least another couple of months. Can we finalize this as Not-a-residence?

Answer:
Yes, code this number as Not-a-residence. Note that we would code it this way even if we were told that someone was going to move in on the next day. The status of a phone number is fixed at the time of the first contact with anyone at that number. And in this case, no one was living in the house at the time of our first contact.
(TP 2/5/96)


Problem: Called a farm

Question:
When we called the sampled phone number, a man said, "This is a farm," and told us that no one lives there. How should this be finalized?

Answer:
If no one lives on the farm, it is like a business and should be coded as Not-a-residence. If someone were living there, however, we would enumerate the eligible persons and select one for an interview.
(TP 8/8/94)


Vacation Homes


Problem: Reached a cottage

Question:
We called the sampled number 12 times, and no one ever answered. Then on the 13th call we reached an elderly man who confirmed the number and said, "You've reached a cottage." The interviewer asked if anyone lived there, and he said, "not year round," and then he hung up. Is this a residence?

Answer:
No, this number should be considered Not-a-Residence. As a general rule, we do not count vacation homes as residences, except for year-round caretakers and so on, so you should not try to interview this person. If we are going to include someone in a sample of households, it must be done through the residence of that person. Otherwise, persons with vacation homes would have an extra chance of falling into the sample. Furthermore, we would have no way of adjusting for that increased probability, since it is not feasible to grill every respondent about possible vacation-home telephones at which they might sometimes be reached.
(TP 2/1/96)


Problem: Call forwarded to vacation location

Question:
Our sampled phone number is the home number of our informant. However, during the summer, it gets forwarded to the "United Methodist Camps," where he spends the entire summer. In other words, we're reaching the household resident at his home phone number but he doesn't happen to be there. How should we deal with this?

Answer:
Conduct the interview if possible. In a random-digit sample, the residential telephone number is the "pipeline" into a household. Once a telephone number has been selected, the residents of the corresponding household can also be selected. Note that if we had selected the phone number of the summer camp, we would not have interviewed the people who have other residences (although year-round caretakers would be eligible).
(TP 8/8/94)


Problem: Summer home for college student

Question:
A male informant said he was renting the place for a couple of months during the summer; during the rest of the year he lives in New Hampshire, where he goes to school. He told me that no one else lives at the place he is renting; furthermore, when I asked him whether he was maintaining a residence at school while he was away, he said, "Technically yes -- I live in a frat house." Should I try to interview him?

Answer:
No, do not attempt an interview. Treat this as a regular vacation home, and code this number as Not-a-Residence. The only way we would select this person into a household sample is if he had his own phone line where he lives when he is at college.
(TP 8/8/94)


Group Residences and Dorms


Problem: Switchboard of retirement home

Question:
We're reaching the "main switchboard" number for a retirement home. Should we code this number Not-a-residence, or should we attempt to enumerate and select someone to interview?

Answer:
Institutions like this are generally excluded from household-based samples, and this number should be coded Not-a-residence. Note, however, that if we reach residents of a retirement home on a personal phone, we would attempt to complete an interview, if the person is otherwise eligible for the study. There are many retirement homes today that are part way between regular apartments and rest homes. If elderly persons in such homes have private phones, that is an indication that they are not simply institutionalized, and we would not want to exclude them unnecessarily from an RDD sample.
(TP 2/5/96)


Problem: Phone in a college dorm

Question:
An informant at a sampled phone number told us that we reached the main number for a college dorm. Do we try to enumerate and select a respondent?

Answer:
No, code the main number for a college dorm as Not-a-residence. However, if we reach college students on a personal phone, we would attempt to complete an interview, if the person is otherwise eligible for the study. (See next question.)
(TP 2/5/96)


Problem: Personal phone in college dorm

Question:
We reached someone at her personal phone in her room at a college dormitory. Is this a group quarters, or should we attempt to do an interview?

Answer:
In a telephone sample, personal phones in college dorms are considered the same as phones in off-campus apartments. This is basically our only chance to include this segment of the population in a telephone sample, and we should attempt to complete an interview. Students in dorms are excluded from face-to-face household samples largely for practical reasons. An advantage of telephone sampling is that we can more easily reach this part of the population.
(TP 2/5/96)


Bad phone numbers


Problem: Reach wrong number

Question:
What do we do when we reach a phone number that is not the sampled number? In these cases an informant tells us that the number we have reached is different from the number we called. Can we code these Not-in-service?

Answer:
Yes, a selected telephone number that rings at some other residential number should be considered Not-in-service. One dialing is not enough to confirm this problem, but if the same thing happens twice, the case can be finaled as Not-in-service.
If the sampled number rings at some other number that is a business, the most likely explanation is that the selected number is part of a seek group at that business; the appropriate final outcome in such cases is Not-a-residence.
(TP 8/8/94)


Problem: Reach different but similar phone number

Question:
Our sampled phone number is (YYY) 123-2320. We're reaching 123-2322. A person who answered told us that there are no other phone numbers that begin with 123-232x but theirs, and that even if we dial 123-2329 we'd still reach their phone number. It sounds like our phone number is a non-existent number and, if so, the proper code is Not-in-service. Can I go ahead and finalize this case as Not-in-service?

Answer:
If this problem is confirmed with a second call, and the number reached is a residence, the sampled phone number should be coded Not-in-service. Otherwise, phone numbers that are not specifically selected would have extra probability of being selected -- their own, plus that of the other numbers that ring at that number. We should only interview persons who were reached through the originally sampled telephone number.
If the number at which the sampled number rings is a business, the most likely explanation is that the sampled number is part of a seek group at that business; the appropriate final outcome in such cases is Not-a-residence.
(TP 2/6/96)


Problem: Phone always busy

Question:
This case was coded "busy" 18 times. An interviewer called it with operator assistance and the operator said that there is no conversation on the line and that it is "probably" out of order. Do we code this Not-in-service or Assumed-not-a-residence (18+ calls never answered)?

Answer:
With operator assistance, we can code this number Not-in-service. This is the cleanest situation, since otherwise we cannot be absolutely sure that it is not just a very busy line (teenager, computer, or whatnot).
Even without operator assistance, however, we can still conclude that that this phone number is Assumed-not-a-residence, since it is not likely that there would be so many busy signals on a true residential line. Any number that is called 18 or more times (spaced over various days and times) and is never answered by a person or an answering machine (with a message), can be coded Assumed-not-a-residence.
(TP 2/13/96)


Problem: Phone off hook

Question:
This number has been called 18 times, and it is always busy. The operator checked and said it was off the hook. What do we do with this?

Answer:
Code this number Assumed-not-a-residence, as we usually do after 18 well-spaced calls that are never answered either by a person or by an answering machine (with a message). It is not likely that there would be so many busy signals on a true residential line. The operator's statement that the phone was off the hook is not to be taken as definitive, since operators often guess about these matters. This is probably just a non-working number.
(TP 2/13/96)


Problem: Temporarily disconnected or circuits busy

Question:
On some cases we get messages like, "the number you have reached has been temporarily disconnected," or "this number is being checked for trouble so please try your call later," or "all circuits are busy now." How should they be finalized?

Answer:
These numbers should be left in the scheduling queue so that they will be tried at different times. Sometimes it really is a temporary problem. Frequently, however, a non-working phone number will generate these apparently temporary messages. If, however, no one has ever answered this number AND if no answering machine (with a message) was encountered, we can code such a number as Not-in-service after 3 calls that result in these messages or in a combination of these messages and funny signals such as fast-busy or beeps.
(TP 2/5/96)


Problem: Funny signals

Question:
We have a number that interviewers have coded variously as "bad ring/fast busy," "busy," or "answering machine." What we're getting is a series of beeps followed by a fast busy signal. Operators have told us that it "sounds like a call forwarded to a non-working answering machine, then routed through a switchboard to a fast busy signal -- probably a business." I think that we really shouldn't rely on what these operators tell us in these situations. This does not seem to be a good number. How many times do we have to try it?

Answer:
Unless we get a message saying that a sampled number is out of service, there is a chance that the problems with a line are only temporary. However, we do not need to call such numbers indefinitely. The key consideration is whether or not anyone has ever answered this phone number OR whether or not an answering machine (with a message) has been encountered. In either of those cases, we will continue to call the number until the situation is resolved. If, however, no one has ever answered and no answering machine (with a message) was encountered, we can code such a number as Not-in-service after 3 calls that result in funny signals.
(TP 2/5/96)


Problem: Beeps combined with no-answers

Question:
We called the sampled number several times. Sometimes we get a beep like an answering machine, but there is no message. Other times the phone just rings. How many times should we keep calling this number?

Answer:
The decision on a phone number like this depends both on the number of times we encounter beeps or other funny signals, and the total number of calls. If no one has ever answered this number and if no answering machine (with a message) was ever encountered, we can code this number as Not-in-service after 3 calls that result in beeps or other funny signals.
Or if we make 18 calls, resulting in a combination of no-answers, busy, and funny signals, the number can be coded Assumed-not-a-residence. (TP 2/13/96)


Problem: Circuit problems combined with busy signals

Question:
At the sampled number we sometimes we get a message indicating circuit problems like, "It is not necessary to dial 1 for this number." Most of the time, however, the line is busy. How many times should we keep calling this number?

Answer:
What to do with a phone number like this depends both on the number of times we get the message indicating circuit problems, and the total number of calls. If no one has ever answered this number and if no answering machine (with a message) was ever encountered, we can code this number as Not-in-service after 3 calls that result in circuit problems, beeps, or other funny signals.
Or if we make 18 calls that result in a combination of busy, no-answers, and funny signals, the phone number can be coded Assumed-not-a-residence.
(TP 2/13/96)


Interviewer errors


Problem: Interviewed wrong person

Question:
After we had completed an interview, the "respondent" revealed that he was only 16 years old and had been playing a little joke (he'd said he was 19.) Should we re-enumerate and attempt to get a "real" interview? If so, do I need to do anything special with the collected data?

Answer:
This case should be redone; otherwise, it will have to be counted as non-response. For CATI studies, however, it can be a big job to re-interview. If you use the same unmodified case, every answer will be a 'change answer'. It will probably be better to have the CATI technical person restore the case to the original setup condition before attempting the new interview.
(TP 8/8/94)


Problem: Miscoded outcome

Question:
Two calls were made to the same number a few hours apart. The first call was coded as an answering-machine. However, since no record was made of the content of the message, as interviewers are instructed to do, it was probably not really an answering machine. The second call was coded Not-in-service/disconnected. It seems unlikely that the number was disconnected in that span of a few hours; more likely, the interviewer on the first call mistakenly coded a funny signal as an answering-machine. How can I finalize this case?

Answer:
Since there seems to be evidence that the original outcome (Answering-machine) was a miscode, it can be ignored for purposes of assigning a final outcome code. If the Not-in-service outcome has been confirmed by a second dialing, this case should be finalized as a phone number Not-in-service.
(TP 2/6/96)