Psychology: Happiness and Productivity, Summaries of Psychology

Download Psychology: Happiness and Productivity and more Summaries Psychology in PDF only on Docsity!

D I S C U S S I O N

P A P E R

S E R I E S

Happiness and Productivity

IZA DP No. 4645

December 2009

Andrew J. Oswald

Eugenio Proto

Daniel Sgroi

IZA Discussion Paper No. 4645 December 2009

ABSTRACT

Happiness and Productivity*

The paper provides evidence that happiness raises productivity. In Experiment 1, a randomized trial is designed. Some subjects have their happiness levels increased, while those in a control group do not. Treated subjects have 12% greater productivity in a paid piece-rate Niederle-Vesterlund task. They alter output but not the per-piece quality of their work. To check the robustness and lasting nature of this kind of effect, a complementary Experiment 2 is designed. In this, major real-world unhappiness shocks – bereavement and family illness – are studied. The findings from (real-life) Experiment 2 match those from (random-assignment) Experiment 1.

JEL Classification: D03, J24, C

Keywords: labor productivity, emotions, well-being, happiness, positive affect, experimental economics

• (^) For fine research assistance, and valuable discussions, we are indebted to Malena Digiuni, Alex Dobson, Stephen Lovelady and Lucy Rippon. For helpful advice, we would like to record our deep gratitude to Alice Isen. Seminar audiences at PSE Paris, the University of York, and the University of Zurich provided insightful suggestions. Thanks also go to Johannes Abeler, Eve Caroli, Emanuele Castano, Andrew Clark, Alain Cohn, Ernst Fehr, Justina Fischer, Bruno Frey, Dan Gilbert, Amanda Goodall, Greg Jones, Graham Loomes, Rocco Macchiavello, Michel Marechal, Sharun Mukand, Daniel Schunk, Claudia Senik, Tania Singer, and Luca Stanca. The first author thanks the University of Zurich for a visiting professorship. The ESRC and the Leverhulme Trust provided research support.

(Isen et al. 1978; Teasdale and Fogarty 1979), and leads to greater altruism (Isen and Simmonds 1978). 1 These findings apply to unpaid settings. The present paper implements an empirical test that has not been performed in the psychology literature. By doing so, we address a question that is of special interest to economists (and perhaps arguably also to economic policy-makers): Does happiness make people more productive in a paid task? The paper finds that it does. We demonstrate this -- using two different setups -- in a piece-rate ‘white-collar’ setting^2 with otherwise well-understood properties.^3 Interestingly, the effect operates through a rise in sheer output rather than in the per- item quality of the laboratory subjects’ work. Effort increases. Precision remains unaltered. In the first part of the paper, we do not distinguish in a sharp way between happiness and mood. For simplicity, we take the distinction, in a short run experiment like the one initially described later, to be predominantly semantic. Nor do we explore the possibility that other stimuli such as music, alcohol or sheer relaxation time -- all mentioned by readers of early drafts -- could have equivalent effects. Nor can we assess exactly how long-lasting are the effects of emotion upon labor productivity. In a second experiment, however, we turn to such issues. Here we draw upon important external shocks from Nature as a real-life source of variation.

In later sections we describe the existing related literature. Section 3 is devoted to a theoretical model that, it is suggested, can provide a useful conceptual framework. Sections 4 and 5 describe the experiment. Section 6 presents the main results and Section 7 some empirical checks. In Section 8, we examine questionnaire responses that shed light on subjects’ self-perception and related issues. Section 9 describes a second experiment -- on large shocks from the real world -- that confirms the results of the first experiment. Further results are in the appendix.

2. Background The links between productivity and human well-being are of interest to many kinds of social scientists. Argyle (1989, 2001) points out that little is understood

(^1) A body of related empirical research by psychologists has existed for some years. We list a number of them in the paper’s references; these include a series of papers in the 1980s, Ashby et al. (1999), Erez and Isen (2002), and the recent work ofHermalin and Isen (2008). A survey is available in Isen (1999). Our study also has links to ideas in the broaden-and-build approach of Frederickson and Joiner (2002) and to the arguments of Lyubomirsky et al. (2005). (^2) Such as Niederle and Vesterlund (2007). (^3) The analysis draws on a kind of mood induction procedure that is uncommon in the economics literature but is more familiar to researchers in social psychology. One exception is Kirchsteiger, Rigotti, and Rustichini (2006) who find a substantial impact inthe context of gift-exchange.

about how life satisfaction affects productivity, but that there is (some) evidence that job satisfaction exhibits modestly positive correlations with measures of worker productivity. Wright and Staw (1998) find a significant and sizeable effect of long term happiness on productivity. More specifically, Boehm and Lyubomirsky (2008) preliminarily define a happy person as someone who frequently experiences positive emotions like joy, satisfaction, contentment, enthusiasm and interest. Then, by drawing on both longitudinal and experimental studies, they show that people of this kind are more likely to be successful in their careers. 4 Together with the works mentioned in the previous section, a number of other papers have been interested in positive affect and performance. Work by Wright and Staw (1998) examines the connections between worker affect and supervisors’ ratings of workers. Depending on the affect measure, the authors find mixed results. Amabile et al. (2005) uncovers evidence that happiness provokes greater creativity. In contrast to our paper’s later argument, Sanna et al. (1996) suggests that those individuals in a negative mood put forth a high level of effort. 5

However, these results are all for unpaid activities in the sense that the laboratory subjects’ marginal wage rate is zero.

A small analytical literature written by economists is relevant to our later empirical findings. Although not directly about happiness, it examines intrinsic motivation -- i.e. motivation based on internal psychological incentive -- as opposed to extrinsic motivation (incentivized payments) normally considered in economics. 6 A paper by Benabou and Tirole (2002) focuses on the interactions between self- deception, malleability of memory, ability and effort. The authors consider the possibility that self-confidence enhances the motivation to act, so their framework is consistent with the idea that there can be a connection between mood and productivity. They develop an economic model of why people value their self-image, and they use this specifically to justify seemingly irrational practices such as handicapping self-performance or the practising of self-deception through selective memory loss. Compte and Postlewaite (2004) extend this line of work, by seeking to identify circumstances in which biased perceptions might increase welfare. The

(^4) 5 See Pugno and Depedri (2009) for an extensive survey of this argument.

(2007).See also Baker et al. (1997), Boehm and Lyubomirsky (2008), Paterson et al. (2004), Steele and Aronson (1995) and Tsai et al. (^6) As described in sources such as Laffont and Tirole (1993).

and v as two different sources of utility to the individual. Let e be the energy the worker devotes to solving tasks at work. Let w be the energy the worker devotes to other things -- to ‘distractions’ from work. Let R be the worker’s psychological resources. Hence (e + w) must be less than or equal to R. We assume that u , the utility from work, depends on both the worker’s earnings and the effort put into solving work problems. Then v is the utility from attending broadly to the remaining aspects in life. For concreteness, we could think of this second activity as a form of ‘worrying’. But it can be viewed as a generalized concern for issues in the worker’s life that need his or her cognitive attention. In a paid-task setting, it might be realistic to think of a person as alternating, during the working day, between concentrating on the work task and being distracted by the rest of his or her life. There is a psychic return from the energy devoted to distraction and worry -- just as there is a return from concentrating on the paid task. Consider an initial happiness shock, h. For the sake of clarity, assume separability between the two kinds of utility going to the individual. People then solve the problem: choose paid-task energy e to Maximize u ( p , e , h , z )+ v ( w , h )subject to R ≥ e + w (1)

where the first-order condition for a maximum in this problem is simply Eu (^) e − vw = 0. (2)

The comparative-static result of particular interest here is the response of productivity, given by work effort e , to a rise in the initial happiness shock, h. Formally, it is determined in a standard way. The sign of de*/dh takes the sign of the cross partial of the maximand, so that: Sign de*/dh takes the sign of Eu (^) eh − vwh. (3)

But without more restrictions, this sign could be positive or negative. A happiness shock could increase or decrease the amount of effort put into the work task. To get some insight into the likely economic outcome, consider simple forms of these functions. Let R be normalized to unity. Assume that the u and v functions are concave and differentiable. This is not strictly necessary, but it leads to natural forms of interior solutions. The analysis is easily generalized. How then might an exogenous happiness perturbation, h , enter a person’s objective function? The additive model has a maximand as follows,

u(.) + v(.) + h and is -- we conjecture -- what most economists would write down when asked to think about exogenous emotions and choice. They would view a happiness shock as a vertical shift upwards in the utility function, so that the worker gets the h happiness shock whether or not he or she subsequently works or instead worries about other things. Therefore, the optimal work effort e* is independent of the happiness shock, h , or, put in other words, happier people are neither more careless nor more productive. Another, and arguably more plausible, form of utility function has a happiness shock operating within a concave structure. Imagine the worker solves Maximize u ( pe + h )+ v ( 1 − e + h ) (4)

which is the assumption that h is a shift variable inside the utility function itself, rather than an additive part of that function. Now the first-order condition is u ′ ( pe + h ) p − v ′( 1 − e + h )= 0. (5)

In this case, the optimal level of energy devoted to solving work problems, e* , does depend on the level of the happiness shock, h : The sign of de*/dh takes the sign of u ′′^ ( pe + h ) p − v ′′( 1 − e + h ).

Its first element is thus negative and its second is positive. By the first-order condition, we can replace the piece rate wage term p by the ratio of the marginal utilities from working and worrying. Hence, after substitution, the sign of the comparative static response of work effort, e, with respect to the size of the happiness shock, h, is greater than or equal to zero as uu ′′′(.)(.) (^) − v v ′′′(.)(.)≥ 0. (6)

These terms can be viewed as versions of the degrees of absolute risk aversion in two domains -- the utility from work and the utility from worrying. If the marginal utility of worry declines quickly enough as energy is transferred from working to worrying, then a positive happiness shock will successfully raise the worker’s chosen productivity, e*. Put intuitively, as the individual become happier and condition (6) holds, that allows him or her to divert attention away from other issues in life and become more

and effort to alter overall productivity. Our control variables came from (i) requiring our subjects to do a brief GMAT MATH-style test (5 multiple choice questions) along similar lines to that of Gneezy and Rustichini (2000), and (ii) obtaining information in a final questionnaire to allow us to construct a measure of subjects’ prior exposure to mathematics. The aim was to control for heterogeneous ability.^10 Some means has to be found of inducing an exogenous rise in happiness. The psychology literature offers evidence that movie clips (through their joint operation as a form of audio and visual stimulus) are a means of doing so. They exogenously alter people’s feelings. Westermann et al. (1996) provides a meta-analysis of methods. We used a 10-minute clip based on composite sketches taken from various comedy routines enacted by a well-known British comedian. In order to ensure that the clip and subjects were well-matched, we restricted the laboratory pool to subjects of an English background, who would likely have been exposed to similar humor before. As is explained later, whether subjects enjoyed the clip turned out to be important to the effects on productivity. While the key treatment involved the use of the clip as compared with a control treatment identical but for the lack of a clip, we also wanted to address the possibility that the time spent watching the movie clip might be an important factor. Hence we also ran a second control treatment using a “placebo” film clip designed to be neutral with regard to mood but to take up the same amount of time as the comedy clip. The placebo film essentially consisted of the appearance of colored lines placed randomly on a screen. Usefully, the data revealed that this placebo clip was not significantly different from showing no clip whatsoever. These results are reported in part 2 of the appendix. Experiment 1 generates random variation in happiness across our laboratory subjects. At a broad level, however, we are interested in whether natural real-world variation in happiness, in response to emotional shocks, might create productivity effects. Moreover, Experiment 1 is intrinsically short-run: we would not expect the impact of the comedy clip to last. Therefore, in a separate set of supplementary experimental sessions, we asked subjects to report significant real-world shocks in the previous few years -- including family health issues and deaths. Section 9 reports our

(^10) We deliberately kept the number of GMAT MATH-style questions low. This was to try to remove any effort component from the task so as to keep it a cleaner measure of raw ability: 5 questions in 5 minutes is a relatively generous amount of time for anIQ-based test, and casual observation indicated that subjects did not have any difficulty giving some answers to the GMAT MATH-style questions, often well within the 5-minute deadline.

findings -- this is Experiment 2 -- in this longer-run context. In summary, the data collected included the number of successful and unsuccessful numerical additions, performance in a brief GMAT MATH-style test, and (for a subset of laboratory subjects) responses to a questionnaire that included questions relating to happiness, personal characteristics and intellectual ability.

5. Design in detail In Experiment 1, we randomly assigned people into two groups: Treatment 0 : the control group who were not exposed to a comedy film clip. Treatment 1 : the treated group who were exposed to the comedy clip. The control-group individuals were never present in the same room with the treated subjects (hence they never overheard laughter, or had any other interaction). The experiment was carried out on six days, with deliberate alteration of the early and late afternoon slots, so as to avoid underlying time-of-day effects, as follows. Our main experiment took place over 4 days and 8 sessions; we then added 4 more sessions to check for the robustness of our central result to both the introduction of an explicit payment and a placebo film (shown to the otherwise untreated group). Accordingly, the experiment consists of - Day 1: session 1 (treatment 0 only), session 2 (treatment 1 only). - Day 2: session 1 (treatment 0 only), session 2 (treatment 1 only). - Day 3: session 1 (treatment 1 only), session 2 (treatment 0 only). - Day 4: session 1 (treatment 1 only), session 2 (treatment 0 only). plus
   • Day 5: session 1 (treatment 1 and explicit payment), session 2 (treatment 0 and placebo clip)
   • Day 6: session 1 (treatment 0 and explicit payment), session 2 (treatment 1 and explicit payment) Subjects were allowed to take part on only one day and in a single session. On arrival in the lab, individuals were randomly allocated an ID, and made aware that the tasks at hand would be completed anonymously. They were asked to refrain from communication with each other. Those in treatment 1 (the Happiness Treatment subjects) were asked to watch a 10 minute comedy clip designed to raise

them beforehand, was 10 minutes. Each subject had a randomly designed sequence of these arithmetical questions. Numerical additions were undertaken directly through a protected Excel spreadsheet, with a typical example as in Legend 1. The spreadsheet necessarily contained more such rows that any subject could hope to add in the ten minutes allowed. The subjects were not allowed to use calculators, and it was explained that any attempt to use a calculator or any outside assistance was deemed to be a disqualification offence, resulting in only the show-up fee being paid, though they were allowed to use pen and paper and these were provided for their use. This did not prove to be a problem across the 4 experimental days. The numerical additions were designed to be reasonably simple, if repetitive, and earlier literature has deemed this a good measure of intellectual effort (Niederle and Vesterlund, 2007).

Legend 1: Adding 2-digit Numbers

The second task for subjects was to complete a simple 5-question GMAT MATH-style test. These questions were provided on paper, and the answers were inputted into a prepared protected Excel spreadsheet. The exact questions are given in an appendix. This test was designed as a brief check on ability, as used before in the research literature (Gneezy and Rustichini, 2000). The final task, which was not subject to a performance-related payment (and subjects were made aware of this), was to complete a questionnaire. A copy of this is provided in an appendix. The questionnaire inquired into both the happiness level of subjects (before and after the clip for treatment 1), and their level of mathematical expertise. The wording was designed to be straightforward to answer; anonymity was once again stressed before it was undertaken; the scale used was a conventional 7- point metric, following the well-being literature. Moreover, in day 5 and day 6, we added extra questions (as detailed in the appendix). These were designed to inquire into subjects’ motivations and their own perceptions of what was happening to them. The purpose was to try to shed light on the psychological mechanism that made our treated subjects work harder.

To summarize the timeline for Experiment 1: 12

_1. Subjects enter and are given basic instructions on experimental etiquette.

2. Subjects in treatment 1 are exposed to a comedy clip for 10 minutes, otherwise not.
3. Subjects are given additional instructions, including a statement that their final payment relates to the number of correct answers, and instructed against the use of calculators or similar devices.
4. Subjects move to their networked consoles and undertake the numerical additions for 10 minutes.
5. Results are saved and a new task is initiated, with subjects undertaking the GMAT MATH-style test for 5 minutes.
6. Results are again saved, and subjects then complete the final questionnaire.
7. After the questionnaire has been completed, subjects receive payment as calculated by the central computer._
8. Principal results from Experiment 1 A group of 276 subjects drawn from the University of Warwick participated in the experiment. Of these, 182 took part in the main experiment, while the others participated in the sessions of day 5 and 6. Each was part of only one session. A breakdown of the numbers per day and per session is contained in Table 1. The subject pool in the main version of Experiment 1 was made up of 110 males and 72 females. Tables 2 and 3 summarize the means and standard deviations of the main variables (respectively for the treated and untreated subjects). The first variable, the key one in our analysis, is the number of correct additions in the allotted ten minutes. ‘Happiness before’ is the self-reported level of happiness for all subjects (before the clip for the treated group) on a seven point scale. The variable ‘happiness after’ is the level of happiness after the clip for the treated group; ‘GMAT MATH’ is the number of correct problems solved in that section; ‘high-school-grades’ is an index calculated from the questionnaire. Enjoyment-of-clip is a measure in a range between 1 and 7 of how much they said they liked the movie clip.

(^12) The full instructions provided in the appendix provide a description of the timing.

‘Day 2’, ‘Day 3’ and ‘Day 4’ are day-of-the-week dummies. Consistent with the result seen in the previous session, the subjects’ performances are higher in the treated sessions. As we can see in Table 7’s regression (1), in the first column, this result holds when we control for subjects’ characteristics and periods. The coefficient of 0.118 implies that the happiness treatment increases people’s productivity by approximately 12%. In regression (2) of Table 7, the performances are increasing in the rise in elicited happiness (for the case of untreated subjects, by definition, Change-in-Happiness=0). This result is still true when we restrict the analysis to the treated subjects alone, as in regression (3). The size of the effect is only slightly smaller (than in column 2 of Table 7) at approximately eight and a half percentage points. Because of the known skewness in human-performance data, it is natural to use a logged variable. Nevertheless, as a rough check, Table 8 (columns (1) and (3)) re-runs the first two regressions of Table 7 with a dependent variable defined on absolute values rather than log values. The variable ‘Treatment’ in column (1) remains large and positive. It remains statistically significant when, in column (2) of Table 8, we exclude the performance outliers (here we drop the two extreme laboratory subjects, with respectively 2 and 43 correct additions). Similarly, the coefficient on the variable Change-in-Happiness is statistically significantly different from zero irrespective of whether or not in Table 8 we retain the two outliers: see regressions 3 and 4. Might the pattern in the data be in part a kind of reciprocity effect? Are these laboratory subjects ‘repaying’, or somehow trying to please, the investigators? Such difficulties are not uncommon in economics experiments. However, we would argue that this issue is not a significant problem here. In our experiment, people get paid more for every addition they solve. That money goes to them, so that, if anything, extra productivity hurts rather than aids the investigators. 14 Alternatively, laboratory subjects might wish to reciprocate the expected payments made by experimenters by doing as the experimenter wishes. To partially address this, we added direct questions to the questionnaires in days 5 and 6, asking

(^14) Notwithstanding this point, even if the mood-induction procedure did enhance productivity in the experiment through some feeling of reciprocity on the part of the subjects, this would not be in contrast with our hypothesis. Subjects became happier afterthe clip, assuming against our previous argument that they wanted to reciprocate by working harder. This feeling would always result through a positive increase ofworrying to working. The difference would be that u(.) , the utility from working. This mechanism would result in a positive transfer of h would act to increase u(.) through reciprocity rather than directly. Note also e from that there is nothing specific of an experimental environment in this mechanism that might work in that way in a real-life setting.

subjects: “Did you try your best when asked to add numbers?”; “If so, why? If not, why not?”; “Did you feel that first observing the video clip made you better or worse at adding up numbers?”; and “Can you say why you believe that?” Among the treated subjects, out of 48 answers only 31% thought the clip had a good effect, for 23% this effect was bad, while 42% felt it was irrelevant. The number of subjects who declared that the clip had beneficial consequences was not statistically larger than the number of subjects who felt the effect likely to be bad ( p =0.22). Furthermore, among the 25 subjects who were shown a placebo film -- discussed later -- the answers were similar (bad 44%, good 24%, indifferent 32%). The difference between subjects who thought that the placebo film had a positive influence is not statistically different from the number of subjects thinking that the real treatment had a positive effect ( p =0.26). All this appears to point towards subjects not being able to assess the impact of the clip, and not being entirely sure whether we as experimenters were using the clip to aid or hinder them. Accordingly, since subjects' own perceptions on the impact of the clip on productivity are incorrect, so it is hard to argue that they first worked out what the experimenter wanted and then went about trying to ensure that the experiment was a success. Another concern might be that the subjects convince themselves that the fact of watching a clip per se might enhance performance. In section 7 we discuss this and find that it seems unlikely to be true. We show that individuals who are treated with a placebo clip do not perform significantly differently from individuals who are untreated (and, if anything, they do slightly worse on the additions task). It appears, therefore, that positive emotion invigorates human beings. Yet the mechanism here is unclear. Does happiness have its effect on labor productivity through greater numbers answered or through greater accuracy of the average answer? This distinction is of interest. It might even be viewed as one between industry and talent -- between the consequences of happiness for pure effort compared to effective skill. To inquire into this, we estimate a different kind of equation. Table 9 takes attempted additions (in log terms) as the dependent variable. The results are similar to the ones in Table 7, where we considered the number of correct additions. Attempted additions rise by slightly more than 9%. Then, in Table 10, we run exactly the same regression as in Table 9 but with a different dependent variable. This is an estimated equation for ‘precision’, namely, the ratio of correct-answers to attempted-answers. Interestingly,

the previous pattern holds in this case as well: treated subjects perform on average better than untreated ones. Table 11 establishes this claim. In regression 1, where this new scheme is interacted with treatment, the variable, although negative, is not significant (p value 0.69) and it is actually positive (but again not significant (p value 0.81) in regression 3 where we consider the attempts as the dependent variable. Interestingly, from regression 5 we can see that the dummy payment is positive (although non significant with p value 0.21) with respect to the precision; this seems to suggest that -- if anything-- an explicit payment scheme increases productivity via precision and not via attempts. The above considerations suggest that the impact of happiness on productivity will not change if the payment is specified. These consistency tests are encouraging. Much remains, nevertheless, to be understood. One puzzle generated by the data is about the nature of the transmission channel from human happiness to people’s labor productivity. The paper’s earlier theoretical framework describes a set of cases in which, as a structural or mathematical matter, the correct empirical prediction emerges. However, further experiments will have to be designed to try to probe the precise transmission mechanism. Another consideration which may be relevant -- we thank Greg Jones for this suggestion -- is that happiness could act to increase cognitive flexibility. In some recent work, this has been proposed in a narrow context, of the perception of local versus global aspects of a visual scene (Baumann and Kuhl, 2005; Tan, Jones and Watson, 2009). The argument is fairly simple. If someone is focusing on local aspects, then positive affect improves processing of global aspects; and if focusing on global aspects, then it encourages local processing. Jones and colleagues have called this "encouraging the perceptual underdog", and it is distinct from previous suggestions about, say, positive affect simply promoting global processing. It seems plausible to hypothesize that happiness could have a similar effect on a broader canvas, where labor productivity benefits from the individual worker being encouraged to try out hitherto neglected strategies. Finally, (see Tables 2 and 3) the result of the GMAT-style test are not significantly different between treated and untreated, (3,5 for the first 3.37 for the second). This suggests that the GMAT-style test (see note 12 for more details) is a good control variable; it is unaffected by the treatment.

8. Subjects’ self-perceptions Towards the end of our early experimental trials, it became clear that the main result was occurring again and again -- appearing significant even in sessions with the

fewest subjects. We therefore decided to attempt to probe in a qualitative way into what might be happening. In the light of 73 questionnaires completed by the subjects on days 5 and 6, we can ask which of the ideas discussed in section 3 are consistent with the subjects’ own perceptions. In general, no laboratory subject declared that the treatment induced greater focus, while 10 percent of the treated subjects found the comedy clip distracting (this is significantly different from 0, with p =0.01). Moreover, it seems that subjects disagree on the effect of the treatment on performance: out of 48 answers, for 31% the effect was positive, for 23% the effect was negative, while 42% felt it to be irrelevant. This seems to reflect the ambiguity of the effect of happiness on productivity implicit in our theoretical model. As shown in section 3, this effect is positive only if condition (6) is satisfied, which might be the cause of the variation in subject responses, though again a lack of self-perception about the ultimate effect of the clip is also equally plausible. While 88% of subjects who think that the effect is positive find the clip relaxing, 45% of subjects who think that the effect is negative find it distracting, and 12% still use the word “relaxing” albeit this time to describe a negative impact. If we interpret a pronouncement of “relaxing” by subjects for which the effect was positive as an indication of some relief from outside worries, and the pronouncement of “distracting” by subjects for which the effect was negative as an indication of an inability to focus on the task in hand -- perhaps even an increased preoccupation with outside worries -- this answer might be again consistent with the theoretical model. We need to add a note of caution, because the ambiguity in subjects’ responses might be indicative of a general inability to perceive the true impact of the clip on their own performance. This is not implausible, because no subject was allowed to take part in more than one session, so there was no frame of reference for the subjects to consider. To consider how good the subjects were at correctly identifying the direction of the effect on their performance, we try another approach. The 15 subjects who declared that the treatment had a positive effect made on average 21.33 correct additions, against the 18.54 of the remaining 33 subjects. This difference is insignificantly different from zero ( p =0.15), although the sample here is small. If we consider only subjects who felt relaxed and thought the effect of the clip was positive, the p value is 0.10. A positive side-effect of subjects' inability to perceive the impact of the clip on their own performance is to lessen any concerns about the so-called 'demand effect' through which laboratory subjects might wish to reciprocate the expected payments