The safe-driving sign study: Does it illustrate foot-in-the-door?

[sue_frantz]

I just finished reading—and rereading—Freedman and Fraser’s famous 1966 Journal of Personality and Social Psychology paper, “Compliance without pressure: The foot-in-the-door technique” (Freedman & Fraser, 1966). This is the safe-driving sign study. The paper doesn’t say what I expected it to say.

First, a heads up. Because this paper was published in 1966, there is some disorienting time travel involved. Freedman and Fraser’s first experiment involved making requests of “156 Palo Alto, California housewives” (p. 197) who were randomly chosen from the telephone directory. For those who aren’t familiar with the terms housewives and telephone directory, Google can help with that. The second experiment—and the more famous of the two—included in their final analysis the results from 105 women (no word on whether they were housewives) and 7 men (househusbands, presumably). In the paper’s discussion, Freedman and Fraser used the universal masculine. I did the math; 97.4% of the two experiments’ participants were women. Their writing about “his” reasoning in the discussion threw me a bit. None of that is a knock against Freedman or Fraser—nor was it unexpected. It’s just a little commentary on the year 1966.

Now let’s get to the unexpected part.

The second experiment they report in their paper is the safe-driving sign study. Every third or fourth home along blocks in different neighborhoods in Palo Alto, California were selected as the participants. Each block was randomly assigned to one of the five conditions of the independent variable. Residents were asked to either (1) put a 3-inch square safe-driving sign in a home or car window, (2) put a 3-inch square “Keep California Beautiful” sign in a home or car window, (3) sign a petition to encourage California’s U.S. Senators to pass legislation that would promote safe driving, (4) sign a petition to the same Senators to pass legislation that would “keep California beautiful,” and (5) a control group who did not receive any of these initial requests. 

Two weeks later, a different set of researchers who were blind to conditions asked the residents if they would agree to have a large (massive, really), poorly-lettered, safe-driving sign placed in their front yard that would apparently obscure most of the front of their house for seven to ten days.

So far so good. Now let’s turn to the results. They write,

First, it should be noted that there were no large differences among the experimental conditions in the percentages of subjects agreeing to the first request. Although somewhat more subjects agreed to post the “Keep California Beautiful” sign and somewhat fewer to sign the beauty petition, none of these differences approach significance.

The important figures are the number of subjects in each group who agreed to the large request… The figures for the four experimental groups include all subjects who were approached the first time, regardless of whether or not they agreed to the small request (p. 200).

Wait. “[R]egardless of whether or not they agreed to the small request.” This is the experiment that is held up as the classic example of foot-in-the-door. In foot-in-the-door, it’s agreement to that initial request that is so important. Some secondary sources report that nearly everyone in the study agreed to the small request. I don’t see that reported in this paper. All I see is that agreement across the groups was pretty much the same. Same large? Same small? Same in between? They don’t tell us. If anyone knows where the secondary sources got that there was large agreement to the initial requests, please let me know.  

In their discussion, Freedman and Fraser write at length about how receiving a small request can lead to later agreement to a larger request. For example, “It is immediately apparent that the first request tended to increase the degree of compliance with the second request” (p. 200), and “regardless of whether or not the two requests are similar in either issue or task, simply having the first request tends to increase the likelihood that the subject with comply with a subsequent, larger request” (p. 201). Nowhere do they say that agreement to that first request is important.

Now the results. Of those who were asked to display the small safe-driving sign (regardless of how many actually agreed), 76% said yes to displaying the large safe-driving sign. For the other three experimental conditions (display small “Keep California Beautiful” sign or signing either of the petitions), 47% said yes to the large safe-driving sign. How many in the control group said yes to the big sign? “[F]ewer than 20%” (p. 200).

While the safe-driving sign is the more famous experiment, the first experiment Freedman and Fraser report in their 1966 paper does indeed illustrate foot-in-the-door, although they don’t quite report the data that I want.

In this experiment, researchers called 156 randomly-selected phone numbers from the “telephone directory.” The “housewives” who answered the phone were told that the person calling was from the “California Consumers’ Group.” The women were randomly assigned to one of three conditions: (1) the caller asked if they would answer some survey questions, then if they agreed proceeded to ask the questions (performance condition), (2) the caller asked if they would answer some survey questions, then if they agreed said that he “was just lining up respondents for the survey and would contact her if needed” (p. 197) (agree-only condition), and (3) the caller described his organization and described the survey, but did not make any request (familiarization condition). (I imagine that the women in this last group must have been more confused than anything.) Three days later, all of the participants were called again and a control group was called for the first time. The request was to allow “five or six men from our staff” into their homes to spend two hours to count and classify their household products.

In the results section, Freedman and Fraser write,

Apparently even the small request was not considered trivial by some of the subjects. Only about two thirds of the subjects in the Performance and Agree-Only conditions agreed to answer the questions about household soaps. It might be noted that none of those who refused the first request later agreed to the large request, although…all subjects who were contacted for the small request are included in the data for those groups.

In the performance condition (answered the survey questions), 52.8% agreed to allow a group of strange men into their homes for two hours to dig through their household products. In the agree-only condition (agreed to answer questions but weren’t actually asked any), 33.3% agreed to the large request. In the familiarization condition (heard about the organization—for no apparent reason), 27.8% agreed to the large request. In the control condition, 22% agreed to the large request.

Since Freedman and Fraser included everyone in their reported data, we don’t know what percentage of those who agreed to the small request also agreed to the large request. However, since they tell us that the participants who said no to the initial request also said no to the larger request, we know that those who said yes to the initial request were the only ones who said yes to the larger request. (Whew!) In other words, the actual percentages Freedman and Fraser give us underreport the degree of compliance to a large request after agreeing to a smaller request, but the relative differences in percentages are meaningful.

I encourage you to read Freedman and Fraser’s 1966 paper.

For those interested in replication, take a look at an attempt to replicate the household products study in Poland and Ukraine (Gamian-Wilk & Dolinski, 2020). This could be a fun class discussion on how 1960s California differs from 2003 Poland and 2013 Ukraine—some of which the authors of this paper discuss. For that matter, discussing how 1960s California differs from 2022 anywhere would be worthwhile. This can help students better understand why some studies may not replicate when done exactly the same way. What woman in the world today would agree to allow five or six strange men into her home for two hours to count household products?! Let alone based solely on a phone call from a stranger?!?! Conceptual replications, in this case, may be more illuminating.

References

Freedman, J. L., & Fraser, S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology, 4(2), 195–202. https://doi.org/10.1037/h0023552

Gamian-Wilk, M., & Dolinski, D. (2020). The foot-in-the-door phenomenon 40 and 50 years later: A direct replication of the original Freedman and Fraser study in Poland and in Ukraine. Psychological Reports, 123(6), 2582–2596. https://doi.org/10.1177/0033294119872208