Ambivalence: meaning and application

You would love to fly somewhere on holiday and at the same time feel guilty about the impact of that trip. That is not indifference but two strong judgements about the same subject pulling in opposite directions. Psychologists call this ambivalence, and it helps explain why people put off choices and respond inconsistently to information.

Ambivalence is holding strong positive and strong negative evaluations of the same subject at the same time, such as a product, a habit or an organisation.

It is an established concept from attitude research. Objective ambivalence concerns the strength of both evaluations, subjective ambivalence how torn someone feels. Ambivalence differs from indifference, even though both end up in the middle of an ordinary scale.

Ambivalence, called attitudinal ambivalence in research, is holding strong positive and strong negative evaluations of the same subject at the same time: a product, a habit, an organisation or an idea. Someone may find a flying holiday relaxing and adventurous, and at the same time harmful to the climate.

Van Harreveld, Nohlen and Schneider describe ambivalence as the extent to which reactions to a subject are evaluatively mixed, with favourable and unfavourable elements (van Harreveld et al., 2015). At the same time means that both evaluations are present at once. Strong means that these are clear judgements, not two lukewarm impressions. Ambivalence therefore says something about how an attitude is built up, not only about its direction.

Attitudes were long measured on a single scale from very negative to very positive. In 1972, Kaplan pointed out that a neutral answer on such a scale can mean two things: indifference or ambivalence (Kaplan, 1972). He proposed modifying the semantic differential, a common rating technique, to measure the positive and negative components separately.

In 1995, Thompson, Zanna and Griffin developed this into an index often called the Griffin formula (Thompson et al., 1995). Ambivalence increases as both components become stronger (intensity) and more similar (similarity). Priester and Petty discuss this index and distinguish objective from subjective ambivalence (Priester & Petty, 1996). In their review, Conner and Sparks treat ambivalence as an aspect of attitude strength (Conner & Sparks, 2002).

Objective ambivalence, also called potential ambivalence, is derived from the positive and negative reactions someone has towards a subject. Subjective ambivalence is the experienced conflict: someone perceives their own judgement as mixed, conflicted or indecisive (Priester & Petty, 1996).

The two are related but not the same. Measures of objective and subjective ambivalence do not always correlate highly (van Harreveld et al., 2015). Someone can hold opposing evaluations without feeling torn at that moment. According to Priester and Petty's gradual threshold model, felt ambivalence rises mainly with the number of conflicting reactions, those on the smaller side, and that increase then levels off.

Van Harreveld, van der Pligt and de Liver linked research on ambivalence with theories of decision making and coping in the MAID model, the model of ambivalence-induced discomfort (van Harreveld et al., 2009). According to their review, ambivalence is particularly unpleasant when a person has to make a choice about the subject. The opposing evaluations then become accessible simultaneously and uncertainty about the consequences arises.

Newby-Clark, McGregor and Zanna found that strong potential ambivalence felt uncomfortable mainly when both evaluations came to mind quickly at the same time, and more so for people with a strong preference for consistency (Newby-Clark et al., 2002). People deal with that discomfort differently: some look for information, others postpone the choice or turn elsewhere.

Research points to four relevant consequences.

1. Delay. In studies reviewed by van Harreveld and colleagues, people with an ambivalent attitude postponed a decision more often, especially when the choice was personally important (van Harreveld et al., 2015).
2. Weaker link between attitude and behaviour. Less ambivalent attitudes predicted later intentions and behaviour better than ambivalent attitudes (Armitage & Conner, 2000).
3. Sensitivity to information. In the same publication, ambivalent attitudes changed more in response to a persuasive message. This processing is not always neutral: ambivalent participants processed messages that matched their position more extensively than messages that opposed it (Clark et al., 2008).
4. Stability. Reviews discuss the expectation that ambivalent attitudes are less stable (Conner & Sparks, 2002). The evidence is mixed: Armitage and Conner found no relationship with stability over time.

These are group findings from specific studies, not predictions for one person.

• Indifference. Both evaluations are weak. On a single scale, indifference and ambivalence both end up in the middle (Kaplan, 1972; van Harreveld et al., 2015); you only see the difference when you measure both components separately.
• Bimodality at group level. A group splits into a mainly positive and a mainly negative camp. That is a pattern in a distribution, not a mixed judgement within one person (article on bimodal clusters in preparation).
• Core conflict in associations. This Neurofactor working concept describes two strong associations with opposite valence around the same choice. Ambivalence is the general scientific concept; a core conflict is a working hypothesis at the level of associations (article in preparation).
• Opportunity cost and defensive behaviour. Opportunity cost is about what you give up by choosing, defensive behaviour about rejection as protection. Ambivalence is about two opposing evaluations of the same subject.

A single scale from very negative to very positive is not enough. Research uses two approaches (van Harreveld et al., 2015).

Objective ambivalence. You ask about the positive and the negative evaluation separately, for example: "Looking only at the positive side, how positive are your thoughts and feelings about flying holidays?" and the same question for the negative side. The Griffin formula combines both: ambivalence = (P + N) / 2 − |P − N|, where P is the positive and N the negative score (Thompson et al., 1995; Priester & Petty, 1996). The result is high when both components are strong and similar.

Subjective ambivalence. You ask directly about the experience. Priester and Petty had participants indicate how mixed, conflicted and indecisive their impression was, and averaged those answers (Priester & Petty, 1996).

Preferably use both, plus open questions or association tasks about what each side contains (see How do you measure associations?). The example wordings are illustrations, not a validated questionnaire.

Faced with ambivalence, it is tempting to reinforce only the positive side, ignoring half of the attitude. The directions below are editorial advice derived from the research in this article; they have not been tested as a strategy.

• Acknowledge both sides. People know the downside themselves; a message that takes it seriously fits their view better.
• Use strong, verifiable information. Ambivalent people often process information more thoroughly.
• Allow for selective attention. Information that contradicts one's own position is sometimes processed less extensively (Clark et al., 2008).
• Do not increase guilt or discomfort. Deliberately increasing discomfort to force a choice is ethically questionable (see Neuromarketing ethics).
• Respect delay. Applying pressure does not resolve the underlying conflict.

1. Define subject and choice context. For which choice is the attitude relevant?
2. Measure both evaluations separately. Use two questions instead of one bipolar scale.
3. Measure felt ambivalence. Add questions about feeling torn and indecisive.
4. Choose an index in advance. Record which formula you use and report it.
5. Analyse per person. Separate ambivalence (both high), indifference (both low) and two camps.
6. Go deeper with open questions. What are both sides about?
7. Test and repeat. Test hypotheses for communication and measure again later.

This example is fictional. A travel provider notices that consumers are interested in long-haul flights but keep putting off booking. People say they would love to travel yet feel guilty about the impact of flying.

Question: is this ambivalence, indifference or two groups that think differently about flying?

Available information: an earlier single-scale questionnaire; the average lies around the midpoint.

Suitable approach: a new measurement asks about the positive and the negative evaluation separately, with questions about feeling torn and open questions. The analysis examines per person whether both components are high, both low, or whether there are two camps.

Possible interpretation: if many people score high on both sides and feel torn, delay is understandable, and a message acknowledging the downside may fit better than one about travel enjoyment alone. If there are mainly two camps, each camp needs its own approach. No results are available yet.

Next step: the provider tests two messages with outcome measures set in advance, without playing on guilt.

• Reading a midpoint score as neutral. It can hide ambivalence, indifference or a divided group.
• Equating objective and subjective. A high formula score does not automatically mean that people feel torn.
• Talking away the negative side. The conflict remains.
• Labelling people. "An ambivalent type" turns an attitude about one subject into a personal trait.
• Assuming causality. An association with delay does not prove that ambivalence is the cause in your situation.

Separate measurements show whether a group holds strong positive and negative evaluations at once and whether people experience this as conflict. You cannot simply conclude that ambivalence causes delay or other behaviour; the studies in this article concern specific subjects and populations.

There are also several indices that combine the components differently (Conner & Sparks, 2002; Priester & Petty, 1996); choose one in advance and state it. Ambivalence depends on context: the same person may respond differently at another moment of choice. And group results are not a prediction for one person.

Ambivalence describes a recognisable but often overlooked pattern: people who strongly value and reject something at once. A single scale shows a midpoint score and misses the conflict. Measure the positive and the negative side separately, ask about the experience and take both sides seriously in communication.