Unpacking insanity defence standards: An experimental study of rationality and control tests in criminal law

The current criminal law governing insanity acknowledges that some criminal defendants are not responsible for their actions due to a lack of rationality, meaning the defendant lacks the capacity to appreciate the nature or wrongfulness of their acts. However, a defence excusing a defendant who had a general ability to understand the nature and wrongfulness of their acts but was incapable (or meaningfully incapable) of resisting an impulse to commit the offence (a ‘control test’) is more controversial (see Morse, 2002; Morse, 2009; Penney, 2012; Redding, 2006 ). Different jurisdictions differ in the test that they use to determine whether a defendant is not guilty by reason of insanity (NGRI). In the United Kingdom, insanity is currently decided based on rationality only (The McNaughten Rule), so only defendants with rationality defects are excused and those whose defence rests on lack of control are deemed ineligible for a NGRI verdict. In the United States, 21 states use the McNaughten Rule (based exclusively on the defendant's judged rationality), 16 states and the District of Columbia use the Model Penal Code, a test based on deficits in both rationality and control, 8 states and the federal system follow an adaptation of the Model Penal Code in which the defence is allowed only for cognitive dysfunction when the defendant is unable to understand the criminality of his conduct, and 6 states have abolished any form of the insanity defence ( Robinson, 2014 ). In addition, since 1982, 12 states have adopted the guilty but mentally ill verdict (GBMI). Ideally, if the jury finds the defendant GBMI, he will be evaluated and treated before returning to prison to finish the sentence. In practice, however, these defendants are typically assigned longer sentences and don’t receive any treatment ( Desmond & Lenz, 2010).

Researchers and legal scholars have pointed out that rationality tests that do not allow any defence based on lack of control have become outdated given our current understanding of neurological basis and psychological conditions that place one at risk for impulse control deficits ( Penney, 2012; Redding, 2006 ). Neurological evidence now provides insight into compulsion and lack of impulse control and highlights not only the neurological basis of lack of control, but also its neuroanatomical distinctiveness from lack of rationality ( Hyman & Malenka, 2001; Penney, 2012 ). Legal scholars give examples of cases where neurological damage, specifically damage to the frontal lobe (for details on the significance of the frontal lobe, see Barth, 2007 ), has led individuals who are seemingly rational to commit horrific crimes (for example, see Carrido, 2011).

However, others are worried that even given its separate etiology from a rationality deficit, allowing defendants to plea on the basis of a deficit in impulse control will result in too many being characterised as NGRI. A test assessing lack of impulse control could logically lead to a wide array of defences based on ‘caused’ behaviour, as Professor Stephen Morse suggests, the ‘XYY defence’, and the ‘rotten social background defence.’ ( Morse, 1995 ). This is highly important in relation to the complex concept of personal responsibility in the criminal law: the law is based on the fact that although our actions may be caused, we are still personally responsible for them ( Vincent, 2010 ). Control tests are particularly controversial because defendants in these cases are rational agents.

In addition, there are practical problems with current standards. Firstly, research suggests that extra legal attitudes are playing a significant role in juror determinations of insanity. In particular, it has been shown that jurors tend to use their own construct of what insanity is rather than the legal definition ( Finkel & Handel, 1989; Skeem & Golding, 2001 ), and that legal attitudes and biases are resulting in inaccurate categorisations of defendants and a failure to follow judges’ instructions ( Peters & Lecci, 2012).

The idea that juror conceptions of and attitudes towards insanity (rather than legal standards) are determining their verdicts is supported by empirical analysis of real legal cases. In one study, Callaghan and colleagues investigated the frequency and rate of insanity pleas and acquittals in eight states based on data from sample counties ( Callaghan, Steadman, Mcgreevy, & Robbins, 1991 ). Looking at the states using either the Model Penal Code (i.e., states that employ a test for rationality and a separate test for control deficits) or McNaughten (rationality test alone), the highest acquittal rate for not guilty by reason of insanity or NGRI (percentage of NGRI pleas that resulted in acquittal) was in Washington (87.36%), a state that uses the McNaughten test with the burden of proof on the defendant. Both states using the Model Penal Code with the burden of proof on the defendant had lower acquittal rates (New York–43.34% and Wisconsin–28.24%). In addition, the number of defendants who made insanity pleas per 100 felony indictments was not consistently higher in the states using the Model Penal Code. Although the highest rate of pleas was in Wisconsin (1.59 per 100), the second highest rate of pleas was in Ohio, a McNaughten state with the burden of proof on the defendant, and the rate of pleas in New York was only 0.3 per 100 felony indictments. Due to multiple other differences between the states, it is impossible to make firm conclusions from comparisons of the Model Penal Code and McNaughten states here. However, the data do suggest that the standard used may not affect the number of pleas, or the rate of acquittals.

These problems with the current insanity defence standards may be compounded by increasing use of neuroscience in NGRI cases ( Gurley & Marcus, 2008; Schweitzer & Saks, 2011 ). Evidence suggests that jurors find neuroscience-based evidence to be more persuasive than psychological evidence or evidence of family history ( Rendell, Huss, & Jensen, 2010; Schweitzer & Saks, 2011 ). This could mean a defendant's liability could come to be determined by the extent to which an abnormality could be detected in their brain rather than any legal standard.

One solution to current problems with the insanity defence would be to ask jurors to make a rationality determination and then a separate control determination. This would have the advantage of forcing jurors to think specifically about any abnormalities that a defendant might have and how they should be categorised (giving less manoeuvrability based on extra legal opinions). Psychological theory suggests this would be advantageous as focus on specific rules requiring detailed and conscious processing is associated with “Type 2” thinking, which is predicted to be more accurate and less biased ( Evans & Stanovich, 2013 ). Using this standard would also mean that rationality and control tests could be treated differently while both being acknowledged as at least partial defences. For example, it would be possible to treat the rationality defence as a full defence but the control defence as a partial defence, recognising the rationality of the agent but also the diagnosable control deficit they suffer from.

This idea is explored in this paper, through what we call the “McNaughten+ standard”. Under this standard jurors make two separate determinations–first jurors make a rationality determination using the McNaughten standard. This is the same as the current McNaughten test and defendants who satisfy the test have a complete defence. When making the rationality determination, jurors know they will go on to make a determination regarding control. Defendants who satisfy the test for lack of control are given a partial defence.

The present study used a sample of 477 undergraduate students to investigate how jurors apply insanity standards based on rationality and control, how they interpret rationality and control standards, and how insanity standards can be utilised to aid accurate and unbiased juror decision-making. We also tested the idea of splitting the insanity defence into separate rationality and control tests to see how this affects juror decision-making.

We tested two current legal tests; the Model Penal Code (MPC) test (which includes a rationality prong and a control prong) and the McNaughten test (which includes just a rationality prong) and our own insanity defence proposal, the McNaughten+ standard. Participants read a vignette describing either a defendant with a rationality defect or a control defect to categorise according to the standard they were given.

Using these standards we tested three hypotheses. Firstly, based on previous research we predicted that the specific current legal standard used would not affect jurors’ decisions ( Finkel & Handel, 1989; Peters & Lecci, 2012; Skeem & Golding, 2001 ). Secondly, we predicted that if we included a rationality test and control test separately, defendants would be categorised more accurately according to whether they had a rationality or control deficit. Finally, we predicted that by manipulating the legal test given, we could change the influences and biases that play into juror decision-making. We predicted that simpler and clearer tests would be more likely to produce results less influenced by biases.