统计
从一个广泛的角度看，统计这个词可以应用在任何可以被量化的信息上，比如，在一定时期内平均温度的变化可以确定全球变暖的现象正在发生。统计这个词还可以使用在更狭窄的地方，通过接触或者对于群体中每个个体进行计数，对于一个群体产生可计量的信息。美国的人口普查就是这种意义上的案例。其他的还有美国参议员的投票记录，酒驾中致命交通事故的百分比，过去一百年移民波动的模式，来自单亲家庭未婚妈妈的百分比，不同种族—民族群体教育和收入水平的比较。
就像伯斯特（Joel Best ）说提到，虽然“我们把统计结果当成我们发现的事实，而不是我们创造的数字……统计不能独立存在”，但是它们是复杂信息的汇总。他解释道，有时候统计错误是有意的，但是更多的是“它们是混乱，能力不足，无知或有意选择，自以为是努力制造出来的数据，它们重申了他们的拥护者们认为公正和正确的原则以及利益，当评价任何统计数据时，最好问3个问题：“谁创造了它？为什么目的创造的？它是如何被创造出来的？”
当评价统计信息时，你还要问，统计数据的来源是什么？来源是可靠的吗？数据有多长时间了？自从那些数据被收集以来有没有重要的因素变化了？
调查
调查是专业人士最常用的工具之一，特别是社会科学。因为通过调查获得的数据是可量化的，调查通常被归类在宽泛的统计资料之列。但是，我们单独考虑他们以凸显一个特别显著的特点：调查通常通过联系获得数据，不能包括群体中的每个个体（所有人口），而是群体的样本代表。调查通过电话联系，邮件，或个人访谈。样本可能是随机的，系统性的（比如，电话目录中每十分之一或者百分之一的人），或者分层（群体内各部分的确切比例；比如，51%女性和49%男性）。
乔尔·贝斯特警告说：“公众对大多数社会问题的态度太复杂了，不能简单地将其分为利弊，也不能用一个调查问题来衡量。”还有，主导调查的人认识到“提问的措辞会影响结果，”还有，如果他们不够诚实，他们会按照有利于他们个人议题的方式构建他们的问题。
评价调查结果时，你要问，样本有代表性吗？也就是，被调查的人群中所有人员是否有相同的机会被选中？问题清晰和没有歧义吗？
他们有没有客观地描述而不是带有偏见？在邮件调查中，是否很多人没有回复？如果这样，没有答复的人和答复的人有多少区别？还有，有没有其他的调查支持这个调查者的发现？
正式观察
有两种观察研究方式。在隔离观察方式中，观察者不介入被研究的个体中。比如，一个小孩心理学家可能到访一个学校的操场观察孩子的行为。在介入的观察中，研究者介入被研究者的行为中。一个人类学家，和一个游牧部落一起生活几个月，一起用餐，参加他们的公共活动，他就是介入的观察者。
评价正式观察，你要问，有没有可能观察者的存在扭曲了被观察者的行为？观察延续了足够的时间，和我们得到的结论匹配？结论是否概括的太广泛？（比如，通过对一个游牧群体的观察可能会概括所有的游牧群体，忽略了其他游牧群体可能在重要方面有不同的事实。）
原文：
STATISTICS
In the broad sense, the term statistics applies to any information that can be quantified, for example, the changes in average temperature over a period of time to determine whether the phenomenon of global warming is occurring.The term statistics may also be used more narrowly to mean quantifiable information about a group that is obtained by contacting, or otherwise accounting for, every individual in the group. The U.S. Census is one example of statistics in this sense. Others are the voting records of U.S. senators, the percentage of automobile fatalities involving drunk driving, the fluctuations in immigration patterns over the past century, the percentage of unwed mothers who come from one-parent homes, and the comparative education and income levels of various racial-ethnic groups.
As Joel Best notes, although “we think of statistics as facts that we discover, not as numbers we create . . . , statistics do not exist independently” but are summaries of complex information. Sometimes statistical errors are intentional, he explains, but more often “they are the result of confusion, incompetence, innumeracy, or selective, self-righteous efforts to produce numbers that reaffirm principles and interests that their advocates consider just and right.” Best recommends asking three questions when evaluating any statistic: “Who created it? For what purpose was it created? How was it created?”1
When evaluating statistical information, ask, as well, What is the source of the statistics? Is the source reliable? How old are the data? Have any important factors changed since the data were collected?
SURVEY
Surveys are among the most common tools used by professionals, particularly in the social sciences. Because the data obtained from surveys are quantifiable, surveys are often included under the broad heading of statistics. However, we are considering them separately to highlight one distinguishing characteristic: Surveys typically obtain data by contacting, not every individual in the group (known as a population), but a representative sample of the group. Surveys are conducted by telephone contact, mail, or personal interview. The sampling may be random, systematic (for example, every tenth or hundredth person in a telephone directory), or stratified (the exact proportion of the component members of the group; for example, 51 percent women and 49 percent men).
“Public attitudes toward most social issues,” Joel Best warns, “are too complex to be classified in simple pros and cons, or to be measured by a single survey question.” Moreover, those who conduct surveys realize that “the way questions are worded affects results,” and, if they are dishonest, they can frame their questions in a way that advances their personal agendas.2
When evaluating a survey, ask, Was the sample truly representative? That is, did all members of the total population surveyed have an equal chance of being selected? Were the questions clear and unambiguous？Were they objectively phrased rather than slanted? In the case of a mailed survey, did a significant number fail to respond? If so, how might nonrespondents differ from respondents? Also, do other surveys corroborate the survey’s findings?
FORMAL OBSERVATION
There are two kinds of formal observational studies. In detached observation the observer does not interact with the individuals being studied. A child psychologist, for example, might visit a school playground and watch how the children behave. In participant observation the researcher is involved in the activity being studied. An anthropologist who lived with a nomadic tribe for a period of months, sharing meals with them and taking part in their communal activities, would be a participant observer.
When evaluating formal observation, ask, Is it likely that the presence of the observer distorted the behavior being observed? Was the observation of sufficient duration to permit the conclusions that were drawn? Do the conclusions overgeneralize? (For example, the observations made of a single nomadic group might be generalized to all nomadic groups, ignoring the fact that other nomadic groups may differ in important ways.)
