19

In one of my college psychology classes, a few years ago, my teacher claimed that first impressions drastically shaped our perceptions. To make his point, he said the following:

At the beginning of a semester, researchers asked a teacher to enter his class for only 30s seconds and then to get out immediately. The researchers then asked the students to describe their impression of the teacher. By the end of the semester, after the students spent over 45 hours with the teacher, they asked the students to describe their impression of the teacher now; there was practically no difference.

I took it at face value then, but looking at it from a more skeptical point of view, that seems a bit far-fetched. I can see first impression having an impact on perception, but this little change after so 45 hours is pretty incredible. It could have been simply a teacher trying to get the attention of his apathetic students.

So, my question is:

  • Has such a study been conducted with a large sample?
  • Have the researchers taken step to differentiate "first impressions shape our perceptions" from "first impressions are a surprisingly good judge of character"?
MrHen
  • 6,113
  • 37
  • 59
Borror0
  • 7,581
  • 6
  • 48
  • 67
  • Strongly depends on how interactive and engaging the lecture was. Without much interaction, how is a perception going to change? – Lagerbaer May 31 '11 at 17:15
  • I suggested an edit to the title to make this sound more on-topic. Let me know if this sort of thing is discouraged. :) – MrHen May 31 '11 at 18:04
  • As a data point of 1, I've certainly had strong negative first impressions of some people, and been diametrically wrong, and the opposite as well. Consider one half-term Alaska governor. – Mike Dunlavey Jun 01 '11 at 01:50
  • @Borror0 Would you mind linking to the study? That would give me an anchor for a "large" sample. Also I know similar studies, but they didn't use the same students for both reports (so they didn't look at the change, but at the accuracy). Maybe you mean those? It's relevant at least. – Ruben Jun 01 '11 at 14:09
  • The saying is a game with words. You may emphasize the impression, which is hard - or maybe not at all - changeable. Or you can emphasize the 'first' - you may later do whatever you want, the **first** impression can't be changed afterwards, you can only give different second, third and so on impressions. I realize that your question is aiming in a different direction, but the saying has a strong impression, and maybe not only to me, so I think it should be spoken out. – user unknown Jun 01 '11 at 14:40
  • @Ruben: The teacher didn't tell us the study's name, and I couldn't find it either. That's why I ask. If I could find the study, I would have simply read it. :P – Borror0 Jun 01 '11 at 15:40
  • @Borror0 I thought so. Is it *possible* that he misrepresented the study and meant the Ambady & Rosenthal one? At least my teachers had often not read the classics that they talked about, if it wasn't their specific research interest. Maybe you'll have more luck with the keyword "thin slices"? – Ruben Jun 01 '11 at 16:26
  • @Ruben: Perhaps. I'm considering emailing the teacher, to see if he remembers what study it was. It's highly possible he misrepresented the study. – Borror0 Jun 01 '11 at 19:21
  • @Borror0: When somebody says "studies have shown" my first thought is "they haven't shown *me*". It's perfectly reasonable to ask for sources. – Mike Dunlavey Jun 03 '11 at 20:25

2 Answers2

4

I did not find the study you mentioned, but I know a very similar classic (371 citations on Google Scholar). Maybe you meant Ambady & Rosenthal (1993).

They used the thin slices paradigm (showing people very small amounts of information/behaviour) and predicted end-of-the-semester student evaluations. The people who saw the short video clips were different from the students.

In the 1st study, consensual [aggregated across 9 student raters (in their first study)] judgments of college teachers' molar nonverbal behavior based on very brief (under 30 s) silent video clips significantly predicted global end-of-semester student evaluations of teachers.

So they significantly predicted them and some of the correlations were quite large (.84 for optimism). Some correlations weren't significant, so this is of course not in the realm of "there was practically no difference" (quoting you), but it's also a different study.

This wasn't such a small sample to begin with (13 teachers, 8-20 students per class, so at least 104) and they varied it too (thinner slices, predicting principal's ratings of high school teachers). The paradigm is used quite often in different variations nowadays. You actually asked about a different study, so I didn't go in very deeply. Maybe you think my answer strays too far from the original question.

Question two "Are first impressions a surprisingly good judge of character?":

Yes, this step has been taken, but I wouldn't call it surprisingly good. Peer ratings, even by friends, aren't usually highly correlated with self-reported personality.

There is a classic meta-analysis about thin slices by Ambady & Rosenthal (1992). They found no increase in accuracy of predictions of various objective outcomes after half a minute (to 5 minutes). It's contradicted Carney, Colvin, & Hall (2007) who found that accuracy increased with exposure time, which they varied from 5s to 5min (peak for extraversion after 5 min clip .55). Extraversion after 20 seconds was .41 correlated with self-report. I don't know if you consider this "surprisingly good".
The results are different for less obvious traits (openness).

Are first impressions unchangeable?

Srivastava, Guglielmo, & Beer (2010) (great names, simply great names) found that stability of perceiver effects (peer ratings in students) increases over time. Across the Big Five: Week 1-2: .37; Week 2-3: .55, Week 3-4: .65.
So, they are certainly not unchangeable.

Ruben
  • 2,813
  • 25
  • 29
  • One of your sentences is unfinished. "You actually asked about a different study, so I'm not gonna" – Borror0 Jun 01 '11 at 15:37
  • @Borror0 I'm sorry, this was a bit rushed. I also confused the *# of citations* for two A&R studies. I corrected that and also included the earlier A&R study to make up for it. – Ruben Jun 01 '11 at 16:39
  • basically u and the quoted links answer the question: **how long does is take to get a STABLE first impression** with testing **in a range of 0,5 to 5 min**. These studies are clearly aimed on **advertings/marketing psychology**. OP speaks of **45 hrs** between first impression and post impression. Thats very big difference and your answer quoting "not that much" is imho quite misleading. But it seems every answer giving "some" truth to a inexact question yields here strong rating rather then critic of question itself. Kind of problematic from a scientific point of view. letmegooglelinksforyou – Werner Schmitt Jun 03 '11 at 18:10
  • @Werner Yes, it was misleading. If I had found more relevant research, I would have used that instead. I removed the offending "not that much" from my answer. I think the question wasn't that imprecise, however, it described a specific design after all. I think questions can be criticized, but you should cite your claims and you should not engage in overly general (and dubious) criticism like doubting whether serious psychologists would address this question. Anchor effects etc. are well-known phenomena and mono-causality wasn't even implied in the question. – Ruben Jun 04 '11 at 19:32
-1

of course NOT

Work/get to know somebody long enough and whatever first impression you had will adapt to this long-view character. This question/statement implicates that one can measure a (first) impression two times, one would call an oxymoron. If you make the questioning utterly superficial and short you will earn these self-fulfilling prophecies result. One can ask

  • How can I improve first impression on something
  • How to change first impression on something

But if impression changes itself, it is not a scientific objective measurable question, as it is a singular event per se. Its like asking people to describe thier impression on a bomb explosion and they only tell you the color keeps the same between 2 time points. Or watching the explosion one time with your eye and second time with an ultra slow mow camera showing you much more details. U cannot variate just any desired parameter, so you can conclude anything from it according how you set up the measurement. Especially if the impression also depens on the state of the measurement apparatus (here the human - prejudice, selective view,... )

If someone makes statements on such multi-factor phenomenon and there isn't at all a differentiation into factors necessary/sufficient but implicating mono-causality first question has to be does the question vive out measurable observables, how was it measured, how has it to be measured.

Anyone posting here a study "verificating" this question should accurately look at these points, I doubt serious psychological scientist to even consider such questions...

edited more detailed answer

Be cautios on general mono-causal statements, especially in psychology and social sciences. Ask yourself, would such a study question really make sense and can it prove a mono-causal connection? "American scientist found out... ", I had to think on this statement.

Here the question itself is quite superficial, what is impression, how deep does it go in a passive teaching course, how deep in a vis a vis discussion, in a dialog, how interactive is the teaching, how active are the pupils. How you define these parameters will change the whole result. Are self-fulfilling prophecies excluded? You know when you watch a movie, you often pretty immediately know by only watching the actors who is the bad, who the good guy. So your impression is based on prejudice.

We did a lot of prof teaching rating at university with lot of questions and non superficial properties like competence, pronounciation, etc... If a psychological test is made by letting describe someone something vocally, it cannot be very methodical & scientific. There are clear criteria how such a test has to be structured, how many questions (also doubled) have to be asked at least to get a statistical significant result. Maybe you generalized the statement of your teacher as your question here too drastically.

Mad Scientist
  • 43,643
  • 20
  • 173
  • 192
Werner Schmitt
  • 951
  • 6
  • 15
  • Your answer is more a comment, it does not actually answer the question. [We generally expect answers to cite their claims here](http://meta.skeptics.stackexchange.com/questions/5/must-all-claims-be-referenced), and this question specifically asks about studies. You just add your opinion here, you don't answer the question. That is not how this site works. And please write in complete words, textspeak has no place here. – Mad Scientist Jun 01 '11 at 10:27
  • @fabian: there has to be made a distinction, if a question is objective answerable, cannot be answered, will not find any answer. I like stackexchange and it works on most topics but with skepticism u cannot devide only into answer/no answer or u will earn here a lot of borderline qestions where people are affirmating the prejudices of each other by rating, just to get an answer. But likely this is a discussion for meta.skept. To me this a very crucial point esp. for skeptical questions, as no one of us is expert for everything – Werner Schmitt Jun 01 '11 at 13:50
  • Please type out whole words, all the "u"s, "ur"s, ... are very annoying to read. I've corrected it this time, but you really should type properly the first time. – Mad Scientist Jun 01 '11 at 14:01
  • sry fabian, i used to & thought we as germans are allowed to use acronyms, as english grammar doesnt use capital letters for substantives, which annoys me much more ;) – Werner Schmitt Jun 03 '11 at 18:13