Instead of choosing one approach, policy analysis is a scientific activity or not, comparing and contrasting different theories by exploring their weaknesses and strengths are chosen in the essay. It is because there is no possibility to argue that whether policy analysis is a scientific activity, or not.

The case-study essay begins by giving different definitions of policy analysis and then it illustrates contrast theories about the scientific value of policy analysis. It tries to identify limitations of scientific policy analysis. Although it aims to evaluate the scientific value of policy analysis while exploring these issues, the case-study essay endorses the critical approach of rational policy analysis.

What is Policy Analysis?Although policy analysis is widely described as an applied social science discipline, which uses multiple methods of inquiry and argument to produce and transform policy-relevant information that may be utilised in political settings to resolve policy problems, there are different approaches to the policy analysis. It is possible to mention five different forms of policy analysis. (Richardson 1969 :27)- Descriptive studies of social policy and administration which focus on the origins, intentions and operation of specific policies in areas such as housing, education, health and social services.- Individual case studies or attempts to devise generalizable but largely descriptive propositions about the nature of public policy- making.

– Studies of policy outputs, which involve complex statistical analysis with a number of variables relating to economic, social and political characteristics of local authorities.- Evaluation studies assessing specific policies in terms of the extent to which their outcomes have achieved the objectives of the policy.- Collection and analysis of data studies which have the specific purpose of aiding a policy decision or advising on the implications of alternative policies.In addition, Parsons supports the view that policy analysis is applied rather than pure, problem oriented, both prescriptive and descriptive. (1995: 29) Parsons makes a distinction between analysis of policy and analysis for policy.

Whereas analysis of policy includes policy determination and policy content; analysis for policy encompasses policy advocacy and information for policy. However, there must be a distinction between policy analysis and policy advocacy. Because as a term policy analysis is most closely associated with explaining the causes and consequences of various policies, while policy advocacy prescribes what policies governments ought to pursue.Wildavsky ,who supports the view that policy analysis is an activity creating problems that can be solved, has offered a different explanation of policy analysis. He argues that “every policy is fashioned of tension between resources and objectives, planning and politics, scepticism and dogma. Solving problems involves temporarily resolving these tensions.” (1979:17) He also makes the point that policy analysis is an applied sub-field whose content can not be determined by disciplinary boundaries, but by whatever appears appropriate to the circumstances of the time and the nature of problem.Can Policy Analysis Be Regarded As a Scientific Activity?Some writers such as Stokey and Zeckhauser, make the point that policy analysis is an exact science, on the other hand there are some authors who accept the limitations of rationality.

For instance, Quade argues that policy analysis can not be an exact science since it concerned with helping ‘a decision-maker make a better choice than he otherwise would have made’ and thus it aims to facilitate an ‘effective manipulation of the real world.’ (Quade 1976: 25) In his book, Analysis for Public Decisions, Quade writes about the decision-making, which is constrained and bounded by interests, values and beliefs.In support of this view that Weimer also argues that policy analysis is a client-oriented, advice relevant to public decisions and informed by social values. To what extent ‘facts’ and ‘values’ can be separated in policy analysis is an important question as social values and beliefs can play an important role in policy analysis.In addition, Wildavsky describes policy analysis as an art and a craft rather than a science.

(1979: 15) Because a policy analyst should be self-critical about his or her own assumptions and methods. Successful policy analyst must have some basic skills like developing strategies for quickly understanding the nature of policy problems and the range of possible solutions, a perspective for putting perceived social problems in context, some technical skills to enable him/her to predict better and to evaluate more confidently the consequences of alternative policies, an understanding of political and organisational behaviour in order to predict the feasibility of adoption and successful implementation of policies. However, the most important thing is that the analyst should be aware of his/her own political, ideological biases, preconceptions and assumptions.Political realists Lindblom, Wildavsky, Hogwood and Gunn developed an alternative theory to rational policy analysis view. They stated that policy analysis is an activity, which can help to raise the level of political debate.

They also support the view that policy analysis is an art and a craft rather than a science. Although decision-makers expect solutions from social scientists, it is not possible for them to produce solutions, as there is no one best way of decision-making. All they can do is to measure the impact of current and past public policies.In addition, the forensics approach supports the view that policy analysis is an activity of subjecting policy arguments to the test of persuasiveness, rather than quasi-scientific proof or rational ‘truth’. It is not possible to consider policy analyst as a scientist measuring or experimenting. A theorist, Martin Rein, ‘most widely known proponent of this approach’ states that policy analyst tell stories which can not be disentangled by any scientific procedure or rational technique.

(1976:15) The most important thing is therefore the framework within which decisions about policy are made and how plausible these frameworks are. On the other hand, William Dunn, another proponent of this approach, argues that all stories are not equal, because some are more logical or more truthful than the others are. He also believes that experimentalism and the model of scientific explanation is not possible for policy analysis (1981:23)The Limitations of Rational Policy AnalysisParsons supports the view that the key techniques of rational analysis include cost-benefit analysis, economic forecasting, financial planning, operational research, systems analysis, social indicators and impact assessment. (1995:399) However, the technical base of policy analysis is weak. Since values are at the centre of policy-making, ‘neutral’ or value-free analysis can never be practicable.For example the benefit-cost analysis has serious problems and limitations.

Lynch in his book ‘Policy Analysis in Public Policy Making’ explains these limitations with a few examples. (1975:78)”…

..a new freeway exchange saved an extra one or two minutes for a commuter. What is the value, if anything, of that extra minute? Some would argue that one or two minutes saved on travel time can not be considered a benefit because people do not think such units of time are valuable. The final benefit-cost ratio is normally open to dispute, as there is no established convention on how and what various costs and benefits should be computed.””.

…for example an analyst can say that upgrading the transit service to poverty areas is a benefit and using some method he can assign a dollar amount to that benefit. However, the amount chosen by the analyst could be challenged. Also, another analyst could argue that such improved service is not a benefit.

” (taken from Hoos 1972:136)Moreover, while measuring benefits or costs it is important whether the policy will be implemented for a short term or a long term. Because some policies experience difficulties at first but it can be successful in the long term. Program, Planning and Budgeting System which is a form of cost-benefit analysis specifies the output of a government program, then measures the cost of achieving this output and learns whether benefits exceed the costs. Although PPBS is successful to measure the direct costs it is very difficult to identify the indirect, symbolic costs. In addition, there are some social outcomes, which do not have any price.

For example it is not possible to set values on good health, finding a cure for cancer, teaching poor children to read and write.Social indicators are developed by some social scientists to show social progress. They can be defined as quantitative data, which focus attention on certain social conditions. However they are not enough to measure units of social wellbeing and are open to criticisms because of problems associated with value judgements. Dye makes the point that in the social indicators movement there is an implicit political elitism- the view that social scientists are the best judges of what is “good” for the people. (1992:366)Apart from physical and biological sciences there are also some limitations which are caused from the design of social science research.

In social experiments it is not possible to control all the factors that go into a real world situation. We can give an interesting example of this important matter. In the New Jersey Graduated Work Incentive Experiment; the Federal Office of Economic Opportunity funded a three-year social experiment involving 1,350 families in New Jersey and Pennsylvania.

The aim of the experiment was to answer some serious questions about the impact of welfare payments or the incentives for poor people to work; such as would a guaranteed family income reduce the incentive to work?The experiment continued for 3 years , but in this period some political events occurred in US. In 1969 President Nixon proposed the Family Assistance Plan to Congress, which promised all families a minimum income of 50 percent of the poverty level and a payment reduction of 50 percent of outside earnings. The results of the OEO experiment had not been explained and the Nixon administration pressured OEO to produce favourable supporting evidence that a guaranteed income at the levels proposed in FAP would not reduce incentives to work among the poor. So the results of the experiment showed that there were no differences in the outside earnings of families receiving guaranteed incomes (experimental group) and those who were not (control group). However later results confirmed the preliminary results, which were produced to assist the FAP bill in Congress. The Rand Corporation reanalyzed the Graduated Work Incentive Experiment and reached different results.

Because one important factor University of Wisconsin researchers failed to consider was changed during this experiment. Six months later after the experiment began, New Jersey charged its state law and offered all families very generous welfare benefits so during the experiment both the “control” group and “experimental ” group were being given equal benefits. Rand study reached different results that guaranteed annual income would reduce willingness to work.(Dye 1992:367)There is another set of problems in policy experimentation. In general people tend to modify their behaviour when they know they are being watched. This is known as ” Howthorne effect”. A new policy can appear more successful than the old one at first because of this ‘effect’.

And different results can be obtained with large-scale nation-wide policies than small-scale experiments.Value ConflictsSince policy analysis deals with very subjective topics and relies upon interpretation of results, it can not be “value-free”. This is the major weakness of the scientific policy analysis. For instance, liberal reform oriented social scientists expect liberal reforms to produce positive results. Thus even professional researchers interpret the results of their analyses differently.Further evidence supports the view that in the physical and biological sciences the temptation to cheat in research is reduced by the great danger of being caught and disgraced. Whereas in social experiments the temptation to recode and fudge the data and reinterpret the results is very great. (Dye 1992:372)A professor -James Q.

Wilson- formulates two general laws about all kinds of social science researches. Wilson’s First Law: All policy interventions in social problems produce the intended effect – if those implementing the policy or their friends carry out the research. Wilson’s Second Law: No policy interventions in social problems produces the intended effect -if the research is carried out by independent third parties, especially those sceptical of the policy. (1970:34)The major criticism of the scientific policy analysis view is that policy analysts can not act as objective technicians; because analysts usually have supervisors who can be thought as clients. These clients who have different policy preferences can be immediate supervisors or higher-ranking officials. Sometimes they can ignore the analysis and want to reach different conclusions, which support their political interests.

So value conflicts can occur. That is why policy analysis sometimes produces unexpected and even embarrassing findings. (For an interesting example of this see, Dye: 8)It is possible to argue that different political interests interpret the findings of policy analysis differently. Social scientists’ values or own purposes play an important role in using these findings. Not only experimental evidence but also case studies support this point of view.The Green Card Policy Of Turkish GovernmentThe purpose of this case study is to assess whether policy analysis is an exact scientific activity or not by evaluating results of the green card policy in a small district in Turkey.

The methods used in the research are to collect primary and secondary data.The green card policy is being implemented since 1992.In Turkey a mixed model has been applied in health sector. In other words health services are not free of charge.

In general social security bodies pay for the services but not all people belong to these bodies. Some of them do not have any health insurance. That is why health sector is very important for politicians.In 1992, before elections, Turkish Government began to implement a new policy.

The target group of this policy was poor people who do not have enough money to afford health services. But a lot of people wanted to get a green card and tried to cheat officials by giving them wrong information about their social state. Therefore, total expend amount exceeded sent total payment and this has led to budget deficits. (This case is illustrated in Table 1.)