Inferring discrimination from a failure to reply to questionnaires

A claimant who brings a discrimination case of any sort is entitled to issue a discrimination questionnaire. This is a formal mechanism for requesting information on a range of issues, including the reasons for particular treatment, specific figures or statistics, details of training and disclosure of documents and policies. Since questionnaires can be issued before proceedings have been lodged, the procedure may also help in assessing the merits of a claim.

There is no statutory obligation on an employer to answer a questionnaire at all. However, if he does not reply, replies late (that is, outside eight weeks) or replies evasively, the relevant statutes provide that an employment tribunal is entitled to draw 'whatever inferences it considers appropriate' from that failure, including an inference that the employer's treatment of the claimant was unlawful discrimination.

The EAT recently considered this procedure in D'Silva v NATFHE [2008] IRLR 412. The claimant brought discrimination claims against his employer. His union initially supported his case but later withdrew support. The claimant then brought separate proceedings against the union, alleging that the withdrawal of support was, itself, discrimination and victimisation.

In his race discrimination questionnaire, he asked the union to provide statistics for the refusal of legal assistance to members of ethnic minorities. The union declined to answer the question and stated that their computer system did not allow them to record that information. They argued that the exercise of going through the files of every person who had applied for legal assistance and collating them with other files which showed the ethnic origin of all the members in question would have been disproportionate. The claimant asked the tribunal to draw an inference from the refusal to provide the information but the tribunal accepted the respondent's explanation and refused to do so. He appealed to the EAT.

In upholding the tribunal's decision, the EAT (Underhill J) said this (paras 37'“8):

It is important to emphasise that failures of the kind complained of are only relevant to the extent that they potentially shed light on the actual discrimination complained of and thus, necessarily, on the 'mental processes' of the decision taker.

The EAT considered that it was unlikely that the person who decided to withdraw legal assistance would also have been involved in the decisions about the computer systems. They go on to comment on a tendency among claimants (and sometimes tribunals) to rely on a failure to answer questions as automatically raising a presumption of discrimination and they emphatically reject what they describe as this 'tick-box exercise'.

It is necessary in each case to consider whether in the particular circumstances of that case the failure in question is capable of constituting evidence supporting the inference that the respondent acted discriminatorily in the manner alleged; and if so whether in the light of any explanation supplied it does in fact justify that inference.

A failure to reply to a questionnaire may be highly reprehensible but it may also have no bearing on the discrimination complained of and be incapable of supporting an inference.

The case is a reminder that the questionnaire procedure must be used in a focussed way.

From the employer's perspective, if there is a good reason why a particular question cannot be answered, it may be sensible to be as specific as possible about that reason within the body of the questionnaire itself.

General assertions that a request for information is 'onerous' or 'unreasonable' will have to be amplified later on and the earlier an explanation is given, the more plausible it will be.

Similarly, if the reason for replying to a questionnaire late '“ or not at all '“ is simple carelessness or incompetence, then it may be better to bite the bullet and admit it early. An unattractive (and non-discriminatory) explanation is better than no explanation at all.

The key point for employees remains: issue the questionnaire and issue it in time (generally within three months of the acts complained of). A surprising number of representatives do not bother with the procedure at all and miss out on an invaluable source of raw data. The questions must be tailored to the specific type of discrimination being complained about. There will always be a need for some broad-brush questions but, if the questions are too sweeping, they can be easily avoided. If this happens, it may be worth suggesting a compromise by narrowing the question; if the respondent challenges the relevance of the question, the relevance should be spelt out in correspondence. In some cases it is best to write to the respondent to point out missing replies, inviting them to fill in the gaps; in others, the better tactic will be to let sleeping dogs lie.

The EAT has previously held that a tribunal should take account of inconsistencies between answers given in the reply to the questionnaire and answers provided in the pleaded defence (Hinks v Riva Systems, EAT unreported, 501/96). It is also worth remembering that the EAT held in Dattani v Chief Constable of West Mercia Police [2005] IRLR 327 that the tribunal's power to draw inferences from evasive replies to questions is not confined to questions posed in the questionnaire. It also extends to questions asked in ordinary correspondence or by way of a request for written answers.