Lecture 8: Group performance/productivity

Group performance & productivity

Learning aims & outcomes:

Familiarity with social facilitation effects, their relation to task characteristics, and alternative explanations

Awareness of Steiner’s proposals regarding factors determining group performance

Appreciation of reasons for performance loss on certain types of task

Group performance and productivity

Do groups out-perform individuals?

Are two (or more) heads better than one?

What we know already 

Asch: majority defeats truth

Sometimes

Minority opinions are not welcome

Social facilitation

Co-performance effect:  a consequence of mere presence

Triplett (1898) performance improvement

Allport (1924) performance decrement

Explaining effects of task complexity

Time to complete task in seconds
Zajonc, Heingartner & Herman, 1969

Social facilitation

Mere presence effects:

Triplett (1898) performance improvement

Allport (1924) performance decrement

Explaining effects of task complexity:

Zajonc (1965) arousal/drive: increase in speed, strength, probability of dominant response

Difficult/complex tasks require non-dominant responses.  Therefore arousal impairs performance

Social facilitation: meta-analysis

Will you perform better in exams with or without others present?

Bond & Titus (1983)

241 experiments; 23,970 participants

Effects are real (facilitated performance of simple well-learned tasks, impaired performance of complex, less well learned tasks)

…but small

Social facilitation and expertise

If the presence of an audience is arousing and arousal reinforces dominant, well-learned responses, what audience effects should be expected when the performer is a relative novice versus an expert?

Whose performance will be most affected and in which direction?

Social facilitation

Effects on novices versus experts:

Paulus & Cornelius (1974) gymnasts

Forgas et al. (1980) squash players

Sokill & Mynatt (1984) basket ball players

In all cases the experts’ performances were adversely affected by audiences; there were no effects on novices

Social facilitation: mere presence

Zajonc (1980): arousal due to social uncertainty – i.e., about audience reaction

Schmitt et al (1986 ) performing before an “inattentive” audience

Suggests concern about evaluation or competition is not important.  Mere presence of another is arousing

The mere presence—arousal link

Research on social support: effect of presence of others in stressful settings

Cacioppo et al (1990): lack of physiological arousal resulting from observation  by others

But: did fin evidence indicating heightened attentiveness

Effect of audience evaluation

Does audience need to be attending to, evaluating performer?

Guerin (1986) uncertainty is greatest when we cannot monitor others’ reactions

•

Social facilitation: evaluation apprehension

Cottrell (1972) arousal due to learned association between presence of others and reward/punishment following from others’ evaluations

Cottrell’s non-observing audience study

No effect when others present cannot observe

Knowing one may be evaluated by others is responsible for “social facilitation” effect

Another look at Schmitt et al. (1966)…

Schmitt et al. 1986

Implications?

Mere presence produces facilitation effect

Evaluation apprehension produces an even greater effect

Not immediately apparent unless one considers only the performances on the well-learned tasks

Schmitt et al. 1986

A further comparison:
Schmitt et al. 1986

Evaluation apprehension

But:  Bond  & Titus meta-analysis:

Amount of evaluation did not affect degree of social facilitation (across studies)

Social faciliation: attention conflict

Sanders  & Baron (1975)

Social facilitation effects arise when there are conflicting demands on attention.

S&B show distraction, from whatever source, has exactly the same effect as the presence of others

Social facilitation: task irrelevant processing

Paulus

Division of attention aids performances that are highly automatized, interferes with performances that require attentional resources

The point is: well learned tasks are more likely to be performed automatically, without conscious attention to them.  Such attention actually interferes with an automatized performance; distraction or divided attention means one is less likely to pay explicit attention to what one is doing in performing the task

Performing as a group

As a very broad generalization:  groups are more effective than individuals, but individuals are more effiicient.

So which should we prefer?

It depends…

Steiner (1972): three factors determine group performance (group potential productivity)

1)Task demands: divisible vs unitary; optimising vs  maximising

The husbands and wives problem

   Three married couples are trying to cross a river. The one available boat can only carry three people at one time. Only the husbands can row. No husband would allow his wife to be in the presence of another man unless he also is present.

What kind of a problem is this?  Divisible or Unitary?

It is unitary; you cannot divide it into bits and have different people do the different bits

Performing as a group

Steiner three factors determine group performance

1)Task demands: divisible vs unitary; optimising vs  maximising 2)Group resources: talents, skills, knowledge, numbers, strength 3)Process by which group interacts to perform the task 4)(add?: urgency)

Re. task urgency

This is not really an intrinsic feature of a task

Rather it refers to circumstance that might lead us to favour effectiveness over efficiency.  For example if you want to invent an atomic bomb before your enemies do, you might well be prepared to accept a lot of inefficiency by having huge numbers of scientists searching for the solution in order to increase the chances of getting the answer quickly

Process by which group interacts to perform the task

a.Disjunctive task – one solution must be selected. (Select the solution proposed by the most proficient member?) b.Conjunctive task – group’s productivity is tied to that of least proficient member c.Additive task – group product is sum (or sometimes average – note: in Hogg & Vaughan is distinguished as a fifth possibility  -- a “compensatory task”) of members’ contributions d.Discretionary task -- Group may combine contributions as they please ● ●

Disjunctive tasks

Laughlin (1988): task types

a.Preference: b.Intellective – easily demonstrated (Eureka problems – e.g., husband-wife) c.Intellective – not easily demonstrated

Group decision making

Hall & Watson (1971) Intellective task,– optimal answer, but not easily demonstrated.  Groups do better than the average of the individuals within them, but seldom exceed, or even match performance of the best individual. Why?

Preference for conflict avoiding strategies, e.g., convergence on consensus (Sherif); majority vote; turn taking

Relational vs socio-cognitive resolutions (Doise & Mugny, 1984; group performance on a Piagetian task)

Imperfect information sharing

Groups may fail to achieve their maximum potential because of imperfections in information sharing

Sharing information in groups

Imagine the following situation:

Five individuals each have some information relative to a choice.

They each know the same three things that favour choice A and each knows one different thing that favours choice B

Individually, therefore, each should favour choice A.

But collectively they have five bits of information favouring B, only three favouring A

Information sharing in groups

•Shared and unshared arguments •Stasser & Titus (1985): Groups are not very good at ensuring non-common arguments are shared •Sharing common arguments will tend to polarise the group, especially if the non-common arguments favour the alternative •(note, therefore, this also applies to polarisation in group decision making)

Disjunctive tasks

Laughlin (1988): Decision rules

1.Majority (or plurality) wins 2.Truth wins (if proposed) 3.Truth wins (if proposed and recognised by others)

As task solution becomes less easy to demonstrate, decision rules seem to shift from 2 to 3 to 1.

Two alternatives for groups

Davis & Restle (1963)

a.hierarchical – the most able, best informed dominate b.Egalitarian – contributions distributed, and valued, equally

Egalitarian approaches are more common, especially in ad hoc groups

Group member characteristics

Husband and wife problem

Shaw (1932):  14% of individuals in her sample could solvethe  problem

Shaw’s explanation for superior performance of groups:– better error checking

But more likely explanation: greater likelihood of group containing one solver.  Therefore, larger groups > better performance

The horse trader problem

A man buys a horse for £20

He sells the horse for £30

He buys it back again for £40

He sells it again for £50

How much does he make (or lose) over the course of these transactions?

Group member characteristics matter

Horse trader problem is unitary, optimizing, disjunctive

Any group containing a solver should be successful, but many still fail because:

The “solver” does not talk much

The “solver” has low status

Recall social influence effects

Additive tasks (and effects of group size)

Ringleman (1913) Pulling on a rope: Average force per person was 63kg

What happens if you add their contributions together?

Additive tasks

What Ringleman got:

2 person group   theoretical 126kg

Actual 118kg (loss: 8kg)

3 person group  theoretical 186kg

Actual 160kg (loss: 24kg)

8 person group  theoretical 504kg

Actual 248kg (loss: 256kg)

 Production loss - Ringelmann

Losses in additive tasks: motivation explanations

Social loafing “my contribution will not be noticed”: the larger the group

Free-riding “my contribution will make no difference”: disjunctive groups – the worst member. Conjunctive– the best member

Avoiding exploitation “I don’t want to be the sucker” – lack of confidence that non-contribution will be penalised -- justice

Additive tasks

Latane, Williams & Harkins (1979) cheering and clapping

Accounting for production loss in additive task

Distinguish:

  motivation loss

  coordination loss

In these tasks each accounted for about half the production loss

Conditions that reduce motivation loss due to social loafing

Karau & Williams review of evidence

Increasing the following will reduce motivation loss:

Identifiability & uniqueness of individual contributions

East of evaluating individual contributions

Members’ task involvement and accountability

Task attractiveness

Strength of identification with group

Brainstorming

An additive task

Maximising rather than optimising (explicit instructions – Osbourn, 1957)

“the average person can think up twice as many ideas when working with a group than when working alone”

Groups of four out-perform individuals two to one

But they don’t out-perform four individuals (i.e., a nominal group of the same size)

Why brainstorming doesn’t work

Diehl & Stroebe (1990) – production blocking

Comacho & Paulus (1995) – evaluation apprehension

Paulus & Dziondolet (1993) – social matching

The Köhler effect

Köhler (1926) compared performance of individuals and pairs

Effect greatest when difference in ability was moderate

Hertel et al (2000) did not replicate this

Messe et al (2002): the critical difference is knowledge of other’s ability

Ringelmann vs  Kohler

They seem to report opposing effects

Why is there a difference?

Is it something to do with the nature of the task?

Is it because social comparison is easier in the Kohler task?

Is it because performers are more readily identifiable with their own performance?

The Köhler effect

Other work indicates:

Social comparison (information about what is an appropriate performance level) does not play a role

Task nature (compensatory) is critical

Identifiability of performer is also critical

Other group effects

Compensating for less capable group member   -- Williams & Karau, 1991

The husbands & wives problem

Solution:  Let us denote the three couples: H1-W1, H2-W2, H3-W3.

H1 rows across with W1, leaves her on the other side and rows back.  Comes back with H2 and H3.  H1 gets out, and H2 plus H3 row back.  They pick up W2 , leaving W3 behind, and H2, W2 and H3 row back.  H2 and H3 get out, H3 rows back again to fetch his own wife, W3, and returns with her.