Personality Psychology (740)

         a blog


Cuperman & Ickes (2009)

Filed under: Uncategorized — drcb @ 6:24 pm
30 Comments to “Cuperman & Ickes (2009)”
  1. Ashley Olivera says:

    Ashley Olivera
    Cuperman & Ickes (2009) Reaction Post

    Upon reading the following study, my first reaction was that the Big Five Personality Model seems to be too broad of a scale to accurately create a universal unit of measurement for personality. I feel that basing participants’ behaviors presented at the time of the experiment and categorizing them into five distinct personality traits aren’t realistic. It is difficult to predict an individual’s personality based on the factor of one’s behavior alone because it is also made up of predictors of a much larger scale. Although, the model does attempt to create a form of measurement to human personality, it has yet to accurately predict specific behavior. Another way in which the Big Five Inventory may not be completely reliable is because it doesn’t take into account participants’ current mood at the time the experiment is conducted. This can be easily altered due to external stimuli such as a family death, health problems, etc,.
    I found it interesting to see the following study utilize the use of partner effects to explore the extent of how suspected behavior of ones’ partner affected ones’ own behavior. I believe preconceived notions from stereotypes can make the study biased and affect participants’ behaviors when it comes to partner effects. For example, the stereotype that views women as being more uptight, whereas men are more relaxed, can give the impression that woman have a higher level of neuroticism. One part of the study found that men showed more signs of being self-conscious than their female partners, because they felt that their partner was feeling uneasy as well. In a way, it seems as if partners were feeding off one another’s energies as a means of connecting with their partners. This made me realize how deeply fundamental and within human nature it is to have a sense of belonging in all aspects of human interaction.

    • Gabriella Robinson says:

      Gabriella Robinson’s “Comment” on Ashley’s post. Your comment that they didn’t take into account current mood is so insightful! I study emotions in my research and didn’t even think to critique that the authors didn’t mention any self-report of current mood pre-post experiment. I wrote my reaction post about the validity of the study and didn’t even think of present mood as a threat to whether the data is valid, a potentially “introverted” or a person low on “agreeableness” may have experienced an adverse event recently. Also your point about bias is also a good catch. Having a scale on perceived discrimination, and a measure of demographics one discriminates against could suggest more into whether it was that person -state – or whether that was their natural reaction in the interaction with the other research participant on the couch- trait- measure of personality.

    • Tiara Newson says:

      I completely agree with all of the factors you mentioned within your post. While reading the article I felt that viewing the interaction between two individuals before partaking in an unusual/awkward situation, such as the one they were in, is not sufficient proof of one’s personality. The participants may have felt uncomfortable, or as you stated, were undergoing other circumstances acting as external stimuli. This was a good experiment, but isnt enough to compare one’s behavior to the results of their test. Maybe veiwing other interactions between participants to compare would appear more reliable.

    • Caroline Nester says:

      I am intrigued by the interaction of mood/personality/behavior that you bring up in your post. I agree that mood should have been a variable measured and controlled for in this study. My suspicion is that mood would have a larger impact on non-verbal and verbal behaviors observed than on self-reported personality. I imagine interpersonal communications and interactions would be impacted more by current mood and by emotional “state” than someone’s self-reported personality or their personal “traits.”

    • Lexi Pritchett says:

      To play devil’s advocate for the sake of discussion, I wonder how much current mood would influence an experiment on personality. Features of personality are associated with the emotional processing of events, affecting an individual’s reactivity to a stressor. For instance, neuroticism is associated with more negative evaluations of daily incidents, and higher reactivity to stressors. Therefore I wonder if someone’s current mood would actually highlight aspects of an individual’s personality that are being measured in the study. But I guess to systematically evaluate that, you’d have to control for prior stressful events a little better.

    • drcb says:

      “it has yet to accurately predict specific behavior”

      The Big 5 model does have good predictive utility, as seen through the findings in Cuperman & Ickes, as well as others. E.g.,

  2. Susie McHugh says:

    My comment will address one strength I appreciated in this study, and one characteristic that I perceived as a limitation. I particularly appreciated the clarity with which outcome measures, that is, interactive behaviors, were operationally defined. That is, behavioral definitions were written in observable, measurable language so that two independent observers would likely have little disagreement in their evaluations. Rather than basing measurements off of subjective interpretations of behavior or of mood, quantifiable characteristics of behavior such as frequency of smiles or head nods were made, or duration of laughing. This increases the reliability of inter-observer agreement (IOA), and thus, improves the robustness of findings. In contrast, if observations were made of subjective or ill-defined behavior, such as observation of an “attitude” or “mood” without further definition, this may leave room for disagreement between observers, and thus could reduce a reader’s ability to trust findings. Low reliability is generally perceived by researchers in psychology as a fundamental threat to measurement. By quantifying behavior in robust terms, though, the authors of this study increased trustworthiness of their findings. This contrasts with the 1993 study by Funder and Sneed (that Cuperman and Ickes, 2009 sought to extend) which based their correlations upon observations of how personality traits predict behaviors such as “speaking fluently,” or “displaying ambition,” which are both comparatively subjective.

    A limitation of the Cuperman and Ickes (2009) study is one that I am basing off of my own amateur speculation, rather than comparison to past research (as this is not an area that I am well-read in). Cuperman and Ickes (2009) drew correlations between personality traits and social interactive behaviors based on a 6-min interaction with a stranger, that occurred in the absence of a specific cue or prompt. That is, the experimenter did not explicitly instruct participants to talk to each other. The authors failed to mention how many dyads subsequently did not interact with one another. While entirely interaction-free periods likely did not occur too often (in that the authors measured subtle non-verbal behaviors, and even the most introverted of participants likely at least shared several instances of mutual eye gaze), it is difficult to imagine that in a mere six minutes, that zero interaction-free periods occurred across the study’s sample. I would have preferred for the authors to have reported the frequency of non-interaction, even if it is in fact, zero. Furthermore, the authors make a point to emphasize the unpleasantness of interactions between two partners wherein both scored low on the Agreeableness trait. It is again difficult for me to imagine that in situations where two “disagreeable” individuals were paired together, that none of these pairs refrained from interaction at all, in the absence of the experimenter instructing them to. My hypothesis is that if “disagreeable” individuals were in fact as uncomfortable interacting with one another as the authors point out, that there may have also been a tendency for some of these pairs to opt out of interaction. To extend this, even some of the most extraverted individuals may refrain from interaction with a stranger in the absence of instruction to do so. While such outliers (i.e., interaction-free periods) may not have occurred, I would have liked to learn the authors’ interpretation of such behavior, as I suspect that it did occur in some minority of participants.

    • Carly Tocco says:

      Susie, I agree with the strength that you mentioned about the authors trying hard to come up with behaviors that are overt and easy for raters to agree upon. Unfortunately, some of the nonverbal dimensions they looked at such as “expressed warmth” was not as clearly laid out in terms of being operationally defined, which I found as a limitation. It is possible that this information just simply wasn’t provided in the paper and hopefully was worked on when training raters. Also, I appreciate your perspective on lack of interaction. I agree that individuals high on extraversion may simply refrain from interacting with a stranger without cueing and I believe it may have occurred as well.

    • Karen Abraham says:

      Susie, I agree with your assessment and was impressed with the way behaviors were operationalized to be observable. Reviewing the inter-rater alphas, they all seemed good with the exception of body posture and orientation, which speaks to the points that both you and Carly make about what kinds of observable behaviors tend to lead to rater agreement.

      Speaking to your point about disagreeables interacting less, my reading of Tables 3 and 4 show a significant (as the study defines p<.025) positive actor effect of number and duration of speaking turns, as well as a positive (p<.05) partner effect of number of speaking turns. I think this provides some evidence to your theory that lower agreeableness would lead to less talking, which doesn't seem to show up in the actor x partner analysis. It would have been nice if they had coded for duration of silence in conversation, as it may have shown up there.

  3. Hande says:

    The article by Cuperman & Ickes was an interesting article to read. I am not familiar with personality research but overall I enjoy the study. However, there were some parts that needed better explanation to my opinion. The interaction partners used the words “forced, awkward and strained” when describing someone who is neurotic. I understand that these are not authors’ words but I believe a better operational definition on how to describe neuroticism should’ve been on place as well as what is considered awkward. I believe even people who are high on “neuroticism” differ among themselves and not every person who is emotionally reactive and anxious can be defined awkward. The participants were also psychology students, I’m surprised that there wasn’t a suspicion that they were being monitored. The effects of social desirability seems like a big limitation every time naturalistic observation is used. Also, as a psychology students they might know each other already or at least have seen each other in different classes. This could possibly effect their interest in talking with each other, depending if they have some idea who their partner was before the study, maybe it would’ve been better if there were questions about if the participants knew each other at all prior to the study. Another thing that is interesting is that the authors found out that people who scored high on agreeableness can get along with those who scored low on agreeableness just as well as with those who scored high. On the other hand, a person who scored low on agreeableness would be able to talk to someone high in agreeableness but not low in agreeableness. It shows that even though a partner in conversation may be argumentative you can still have a conversation as long as one person is agreeable. Additionally, it would be interesting to see different combinations of traits have had on interactions. While a person may associate more strongly with on trait on Big 5, I don’t think a single trait can really capture someone’s complete personality. Personally, I don’t use only one trait to describe myself.

    • Ashley Olivera says:

      I completely agree with your statement. I found it interesting how you took into account the possible likelihood of participants’ knowing each other prior to the experiment. I believe that would have a significant effect on the results because participants who have some sort of history will have a more relaxed and enjoyable interaction than those who are just meeting for the first time. The setting itself, may have also had some effect on the “awkwardness” of the participants’ interactions. If the participants were to have met for the first time among a group of peers in a social gathering, the interaction might have been more relaxed in nature because there wasn’t an experimental setting to their interaction.

    • Melissa Henao says:

      Hande, I agree with a lot of what you said, because I too, am not too familiar with personality research. Especially when it came to how they described a neurotic person– I also felt that the words used to describe neuroticism were somewhat off to what I would think a neurotic person is. When being put in a room with a sofa and knowing you are part of a psych experiment, anyone could act/feel awkward or feel that they have to force a conversation, or opposite (try to not act awkward). I also like that you mentioned the possibility of them knowing each other. That was something I did not think about because I personally did not know anyone from our class, however, I did notice that many of our classmates knew each prior to the class. I think this would definitely alter the results. For example, I might have been paired with someone who did not know me. My partner may normally be very talkative and have no problem sharing information about themselves — scoring a higher level of extroversion on the BFI test they took before sitting with me. But I’m sure how they act with me versus how they would act with someone they already know from other several classes, I think would be very different. I think several factors used for this study could have skewed some of the results. Hypothetically, if the students really had no idea who each others was because it was an intro class, I do not think this study would give the same results if done among different groups of participants with similar settings.

    • drcb says:

      “better operational definition on how to describe neuroticism”

      To be clear, neuroticism was defined according to subjects’ BFI scores. Words like “awkward, strained” were provided by the interaction partners spontaneously.

  4. Mahathi Kosuri says:

    The present study was the first I have read that attempted to measure the Big Five personality dimensions by objectively coding behaviors, as defined by Cuperman and Ickes (2009). Though I did appreciate the methodology and the concept behind defining and operationalizing behaviors and correlating them to the Big Five personality traits, I found the rational behind this methodology somewhat underwhelming.
    Though Cuperman and Ickes (2009) attempted to dichotomize nonverbal and verbal behaviors, I found the way they coded these seven verbal behaviors and nine nonverbal behaviors to be limiting (both in definition and the article’s attempt to explain the reasoning behind their method). I am curious to know which of these nonverbal and verbal behaviors have been proven to be defined by the Big Five personality traits. For example, does one with more extrovert tendencies ask more questions? Or does the number of “verbal acknowledgement” phrases necessarily indicate that someone has a high Agreeableness factor? These disparities also existed when reading about how they coded nonverbal cues. For example, the interpersonal distance and mutual gazes between participants could have been significantly affected several extraneous factors, not to mention could have been very much influenced by cultural factors.
    This brings attention to a significant limitation in this study, which I have noticed is often overlooked when trying to assess personality in an objective light. The study did not mention the participants’ or dyads’ ethnic, racial, or cultural background. Abundant research has made conclusions on the impact of a person’s culture and race on their personality. Though it has been recognized that the Big Five personality traits are seen across racial and ethnic groups, the study does not even mention the racial and ethnic makeup of the dyads observed when reporting the sample characteristics. For example, I would predict a dyad’s interaction, and/or an individual’s extroversion tendencies would be positively correlated with same-race dyads as opposed to interracial dyads (or vice versa perhaps?). When considering how personality traits predict behavior, I believe it is imperative to at least consider the racial and ethnic make-up of the sample.

    • Hande says:

      Mahatri, I agreed with the limitation you mentioned on your post. I also thought about the cultural factors especially in regards to nonverbal cues. There are so many differences when it comes individual’s ethnic & cultural background. We know they used Psych 101 students but we never get to learn how many of them were international students since that could’ve had a substantial affect on results. Having said that, it might not have been their main focus but with this kind of research it seems rather important.

      • Pashminder Kaur says:

        I agree with the limitation that Mahatri and Hande discussed in regards to cultural and ethnics differences. Cuperman and Ickes (2009) didn’t mention the participants’ cultural background in the participant section, which can definitely play a big role towards the result section. Not all participants are similar, as well as their personality. As Hande said, “there are so many differences when it comes to individual’s ethic and cultural background.” Coming from an Asian background and American background; they are quite different. Speaking from an Asian perspective, women are more introverts; whereas, from an American perspective, women are more extraverts. Having that said, cultural difference plays a big role, especially in this case of study.

    • Safa Shehab says:

      I had the same reaction regarding this study’s analysis of behaviors associated with the 5 personality factors. This study’s attempt to relate observed behaviors to the different personality traits (particularly Extraversion and Agreeableness) that are already shown in the literature to be exactly what defines them seemed circular to me (for example, we know that the 6 facets of the trait Extraversion include talkativeness, friendliness, assertiveness, etc.). Also, this study concluded that, as the authors predicted, the 3 others factors did not contribute as much as Extraversion and Agreeableness to dyadic interaction. However, this finding seems to be biased because the authors examined and rated mostly those behaviors associated with their 2 factors of interest, and I suspect that if they rated more of the traits/behaviors associated with the other factors (for example, Openness), they may have found other effects.
      I also agree with your comment on ethnic and cultural differences between the dyads, and would like to add that I also would have liked to know the gender of the raters of the interactions, as this might have had an effect on how some of the behaviors were rated (particularly, for example, observations such as the actors body posture and orientation could have been rated differently if the rater was female or male, and had lower interrater reliabilities than other observational measures).

    • Emnet says:

      Thank you for your post Mahathi, I certainly agree with you. I think when we are investigating personality traits across cultures, the difficult question to address is whether personality trait scales possess conceptual and functional equivalence across cultures. They went into great detailed about outcome measures and how they were operationally defined. Simply including a detailed demographics table in this study would identifying their participants and provide helpful information into how generalizable these findings are. Not only to address racial and ethnic makeup but also age as well.

  5. Gabriella Robinson says:

    Gabriella Robinson reaction post to Cuperman & Ickes (2009):
    Upon reading the experimental paradigm of the Cuperman & Ickes (2009) paper, it brought me to thinking of when I worked on a study that too included deception in its experimental paradigm. I was involved in a study that sought to illicit shame in interpersonal trauma survivors. While the informed consent mentioned shame and that we would be recording them, what could be an obvious attempt of our behavioral paradigm to illicit shame (by not giving them emotional support- not allowed to smile or speak- as they talked about an experience that they had ranked as their most shameful experience in life) most participants said they were unaware of my motives to get them to behave/feel that emotion, when asked in debriefing at the end of the study whether they knew I was “acting.” The part of the Cuperman & Ickes (2009) article that mentions the experimenter was acting (i.e. began searching through a folder of papers and leaves to go look for those papers in efforts to evoke the desired action/being the experiment) makes me curious as to what was in the informed consent that the participants read over before participating, approved by the UT-Arlington Institutional Review Board. Did the consent mention “interactions with others” or “socialness?” As the article reports only 2 of 90 dyads tested had suspicion of being recorded. Which is in line with the low number of people who suspecting me to be acting in my aforementioned study involvement. Given the participants did not suspect/were behaving in their natural manner suggest the data is likely valid, it is surprising the successfulness of deception studies, that people do not suspect despite “informed” consent. It is also curious that they received IRB approval to record people without their consent (given they mention they gave participants consent from after having recorded and subsequently debriefed them).
    Another part of the experimental design reminded me of a study I did with self-report questionnaires of emotion regulation strategies, upon completion of the self-report measures, I then took a behavioral/physiological measure of emotion dysregulation. My study saw a mismatch in self-report and biological/behavioral measures of emotional dysregulation (i.e. what one rated oneself as on the measures did not match that which was considered “objective” biological measure of regulation of hand temperature). We thought this mismatch could be due to the objective nature showing that “truly”/in actuality what they thought (in self-report measures to be successful) was not actually a successful emotion regulation strategy. I find it remarkable that in this experiment Cuperman & Ickes (2009) gained results that saw a match between “objective” behavioral measures of interaction and self-report measures of personality. Although this makes me take a closer look at the constructs being used/operationalized. While the constructs appeared to have good convergent validity, it would be interesting to have a report of how the participants felt about the interaction; did they feel as though they acted as “their normal self” in that interaction. As in my lab we had a post-debriefing questionnaire to see if they still ranked that as their most shameful experience, and felt the experimental design evoked shame (we had them rate their shame felt on a scale as another measure of successful effect of the experiment’s manipulation/attempt to evoke a shameful response). It would be interesting to know if they thought that couch session captured their personality, and why/why not. Overall, my reaction is of admiration for the experimenter’s experimental design. I think it paid good attention to detail as evident in the three-time point measures of distance between participants. In attempts to provide predictive utility of personality research outcomes, I think this article is successful.

    • Mahathi Kosuri says:

      Gabby, the point you made about informed consent is very interested. I didn’t think to consider the logistics when studies use deception. Im also curious about how the informed consent was worded. So in the study you were a part of, your numbers were similar (2:90 ratio)? Thats so surprising!
      I also had the same sentiment when reading about Cuperman & Ickes (2009) finding results that matched “objective” behavioral measures and self-report measures. But for some reason, I am still skeptical about the experiments design and felt there were a few things left out of the methodology. Thanks for your post!

    • Aditya Kulkarni says:

      Hey Gabriella,

      I like your skepticism on the seemingly “perfect” match between the behavioral measures of interaction and the participants’ self-report indications. It would be highly useful to have some distribution data of how much baseline level of each of the traits were present before conducting the analysis, because by what I saw, there was little indication that any dyad was truly “silent”—an ecological finding that would match real life scenarios. Another point to consider, is the short amount of time: 6 minutes. As you said, “they acted as their normal self” but to let the traits take over, may have a latency involved (well, barring any usage of a modern day distractor like a mobile phone for instance). Whether or not their personality was captured in this transient time would be questionable, but I appreciate your insight into how detailed the time point measures were for the participants even if in a short time period.

      P.S. Your position on informed consent and levels of deception was wonderful to read and a true point to consider for any researcher–your participants may try to make you feel good that everything’s okay with your study.

    • drcb says:

      Regarding the IRB stuff, deception is a component of many (most!) social psych experiments. There are some ways to deal with that when it comes to IRB concerns – like here subjects were given the option to delete their videos after learning they were taped. Also, a full and thorough debriefing must always be given at the end of these types of studies.

  6. Carly Tocco says:

    This paper stated that extraversion and agreeableness are most associated to “social behavior,” but I can also imagine that neuroticism and openness to experiences would also be huge for social behavior. For example, if my mind isn’t open to meeting new types of people I may come off differently in terms of sociability, when in fact the motivation behind withdrawing is quite different. Thinking back to in class discussion about what we can gather in the first 30 seconds from someone’s personality, it was interesting to see what we gathered as a class such as “tone of voice” be picked up in the study under “speaks in a loud voice” as a characterization of extraversion. With that said, as a class we missed outward expressions such as “laugh a lot” which were included in the observer ratings in the study. I can see how laughing is a sign of agreeableness and is almost seen as normative within a brief conversation with a person in order for one to determine the conversation as positive. Being in a lab personally and having judged interpersonal interactions, I know just how difficult it can be to objective measure “expressing warmth” which is a behavior that was looked at in this study and a study in my current lab. As a research coder it is much easier to tally the number of times a person uses a “personal pronoun” than it is to tally how many times and how much “warmth” is expressed. I wish the paper operationally defined what they meant by “expressing warmth” because that could range from simply paying attention to smiling after the other person responds. Similarly, having been a confederate or “actor” in a study before, it was pleasant to see research being done to take into account the confound of the actor’s own personality on the effects of the Big 5 in their partner. This paper along with its predecessors (ex: Funder & Sneed, 1993) pave the way for future studies that include actors by displaying the importance of including actor personality attributes on the overall experience of the participant.

    • Daniel Saldana says:

      Carly, I agree with your statement in that other aspects of the Big Five could play a role in “social behavior.” Specifically, it reminds me of an article I read on the communicative function of sad facial expressions (Reed & Descioli, 2017) that could, in part, suggest that aspects of neuroticism (e.g., depressive qualities) could in fact play a major role in communicating with others. I also agree with the need to operationally define the construct of “expressing warmth,” specifically because it does not lend itself to critique. For example, the researchers may have categorized certain behavior under “expressing warmth” when other researchers may have categorized the behavior differently.

      • drcb says:

        The authors say they had “a team of” coders but are unclear as to how many. In any case, in research where behavior is being coded, you’d want to look at the inter-rater reliability to make sure that the behavior is being coded consistently between the research team.

    • Mahathi Kosuri says:

      Carly, I had the same thoughts when reading the article that you had (though I am not in a lab that assess such behaviors objectively). I thought the method was unique and strong but I also would have liked to see more details about how they operationally defined such hard-to-define behaviors (e.g. “expressing warmth”). Im also curious about how they related/ backed up with research the use of “personal pronoun” to “expressing warmth”. Thanks for your post!

      • drcb says:

        Often terms like “warmth” are open to coders’ intuition based on life experience and knowledge, but there must be agreement between the raters for the outcomes to be meaningful.

  7. Katie Dana says:

    Carly, I had similar thoughts when reading the paper. I agree that neuroticism plays a major role in social behavior, and I was surprised that there was not more of a focus on neurotoicism relative to other “Big 5” personality traits. In theory, a person could be extroverted and open, but at the same time socially anxious in certain environments. I wonder if people who are high in neurotocism would be mistakenly coded as low on openness or extroversion based on perceived “withdrawn” social behavior. Further, I imagine a person would be more likely to be coded as being “not expressing warmth” if they appear anxious. It would be interesting to see a study with similar methods that specifically addressed this issue.

  8. drcb says:

    About the question that came up in class – if getting divorced changes one’s personality. Allemand, Hill & Lehmann (2015) found it does somewhat. People who divorced decreased in positive affect and extraversion, and conscientiousness went down a bit too.

Leave a Reply




Spam prevention powered by Akismet

Skip to toolbar