我試著翻譯一些，大家一起來吧!
Study 1: A German study (Stang et al., 2001) compared 118 patients with a rare form of eye cancer called uveal melanoma to 475healthy patients who did not have the eye cancer. The patients' cell phone use was measured using a questionnaire. The eye cancer patients used cell phones more often, on the average.
比較118個眼睛得癌症及475個沒得癌症的病人(以問卷形式)，得癌症者，平均來說比較常用手機。
Study 2: A U.S. study (Muscat et al., 2000) compared 469 patients with brain cancer to 422 patients who did not have brain cancer. The patients' cell-phone use was measured using a questionnaire. The two groups' use of cell phones was similar.
比較了469為腦癌病人及422為非腦癌病人(以問卷形式)，發現二者並無使用頻率上的差別。
Study 3: An Australian study (Repacholi, 1997) conducted an experiment with 200 transgenic mice, specially bred to be susceptible to cancers of the immune system. One hundred mice were exposed for two half-hour periods a day to the same kind of microwaves with roughly the same power as the kind transmitted from a cell phone. The other 100 mice were not exposed. After 18 months, the brain tumor rate for the mice exposed to cell phone radiation was twice as high as the brain tumor rate for the unexposed mice.
將200隻基因轉移老鼠分成兩組，一組每天照射兩個半小時類似手機發出的微波，另一組則無。18個月後，照射微波那組的腦腫瘤比未照射哪組大了兩倍。
Why is it that results of different medical studies sometimes disagree?
為何各醫學研究的結果有時候會不同呢？
What is the best study design for gathering data to explore whether there is an association between cell phone use and the occurrence of cancer? Can such a study design establish a direct causal link between cell phone use and cancer? Is this study design ethically feasible to conduct?
怎樣的研究設計對二者的相關性才是最好的收集資料方法呢？
有研究設計可以證明二者的因果連結嗎？
這樣的設計在倫理上可行嗎?
posted by atx4598 @ 6:15 PM
7 Comments:
A knowledge of different study designs for gathering data helps us understand how contradictory results can happen in scientific research studies and helps us determine which studies deserve our trust.
不同收集資料收集的研究方法，可以讓我們理解在科學研究中矛盾的結果是如何發生；也讓我們知道，我們知道該決定何者值得我們信任。
In this chapter, we'll learn that the study design can have a major impact on its results. Unless the study is well designed and implemented, the results may be meaningless, or worse yet, misleading.
我們將學到，研究設計對研究結果有重大的影響。除非研究有好的設計及工具，否則，結果將是沒有意義的；更有甚者，可能是一種誤導。
For these and other statistical analyses to be useful, we must have "good data." But what's the best way to gather data to ensure they are "good"?
前幾張學到的統計方法都必須建立在「好的數據」上，那什麼方法收集到的才能讓它們是「好的」呢？
The first two studies merely observed subjects, whereas the third study conducted an experiment with them.
我們所舉的三個例子中，前二者只是「觀察」受測者，第三個則是運用了「實驗」。
We'll see both good ways and bad ways of using them. This will help us understand when we can trust the conclusions of a study and when we need to be skeptical.
使用二者各有其優缺點，了解其中的意義可以幫助我們知道，何時該相信/何時該懷疑研究的結果。
We use statistics to learn about a population.
我們以統計去理解母群體。
Since it would be too costly and time consuming to seek out all individuals in the populations of Germany or the United States, the researchers used samples to gather data. These studies, like many, have two variables of primary interest-a response variable and an explanatory variable. The response variable is whether or not a subject has cancer (eye or brain).The explanatory variable is the amount of cell phone use.
既然研究母體全部既費時又費錢，研究者便使用樣本來收集資料。這三個研究中(與其他研究一樣)，有兩個我們感興趣的變項：回應變項(依變相)—受測者是否得癌症；解釋變項(自變項)—使用手機的總量。
Experiment 實驗
A researcher conducts an experiment by assigning subjects to certain experimental conditions and then observing outcomes on the response variable. The experimental conditions, which correspond to assigned values of the explanatory variable, are called treatments.
在實驗過程中，實驗者將受測者安排到特定的實驗條件中，並從其中獲得依變相的結果。符合自變項分配值的實驗條件稱之「處遇」。
For example, the Australian study used mice as the subjects. Each mouse was assigned to one of two treatments, receiving radiation or not receiving radiation. These are the categories of the explanatory variable-whether or not the subject is exposed to cell phone radiation.
以澳大利亞的老鼠實驗為例，所有的老鼠被分成兩個處遇—接受輻射/未接受輻射。
An experiment assigns each subject to a treatment and then observes the response. By contrast, many studies merely observe the values on the response variable and the explanatory variable for the sampled subjects without doing anything to them. Such studies are called observational studies.
實驗分配給每個受試者一個處遇，並觀察其回應。相對的，很多研究只是觀察樣本中自變項及依變相的值，並未對其做任何事。這樣的研究稱之為觀察研究。
作業三的3.4翻譯~(若翻譯錯誤，請見諒~)
3.4
in a recent General Social Survey,respondents answered the question, "In the past month, about how many hours have you spent praying, meditating,reading religious books,listening to religious broadcasts,etc?"The responses on this variable were cross-tabulated with the respondent`s gender.The table shows the results.
在一次最近一般社會勘測, 應答者在上個月回答了問題, ", 關於多少個小時有您被花費的祈禱, 思考, 讀宗教書, 傾聽宗教廣播, etc?" 反應在這可變物十字架被製成表了以respondent`s gender.桌展示結果。
a. identify the response variable and the explanatory variable.
辨認反應易變和說明可變物。
b. find the conditional proportions on the response variable,and interpret them.
發現有條件比例在反應可變物, 和解釋他們。
c. using software,create a side-by-side bar graph that compares males and females on the response variable. summarize results.
使用軟體, 肩並肩創造比較男性並且女性在反應可變物總結結果的一張條線圖。
d. what would be a disadvantage of using two pie charts to make this comparison?
什麼會是使用二個圓形統計圖表不利做這比較?
a Were the German and U.S. studies experiments or observational studies?
那麼，美國及德國的研究是屬於實驗還是觀察研究呢？
In both the German and the U.S. studies, information on the amount of cell phone use was gathered by giving a questionnaire to the sampled subjects. The subjects (people) decided how much they would use a cell phone and, thus, their amount of radiation exposure. The studies merely observed this exposure. No experiment was performed. The studies were therefore observational studies
德國及美國的研究，資料收集都從樣本做問卷。由受測者來決定(回答)他們使用多久的手機，因此，他們受微坡影響的時間亦是由他們決定。這兩個研究只是觀察受測者，當然是觀察研究。
b. How were these studies fundamentally different from the Australian study?
這兩個研究基本與澳洲研究有何差別？
In the Australian study, the researcher rather than the subject (the mouse) decided how much radiation each subject would receive. The researchers did not merely observe the subjects but determined which treatment each would receive. The study was therefore an experimental study rather than an observational study.
澳洲的研究者決定了每個受測者(老鼠)微波所接受的量。他們不只是觀察而且決定了每個受測者的處遇。所以，是一種實驗研究。
One reason that results of different medical studies sometimes disagree is that they are not the same type of study.
以上三個研究之所以有不同的結果，有個原因是：他們採用了不同的研究型態。
“Student drug testing not effective in reducing drug use”
The study found that drug use was similar in schools that tested for drugs and in schools that did not test for drugs.
學生的毒品測試並無法有效地降低毒品的使用
研究發現，不管是否有進行毒品測試，毒品的使用情形是相近的。
a. What were the response and explanatory variables?
The purpose of the study was to compare the percentage of students who used drugs in schools that tested for drugs and in schools that did not test for drugs.
依變相、自變項各為何？
此研究的目的在比較進行毒品測試與沒有進行毒品測試的學校中，學生使用毒品的比例。
b. Was this an observational study, or was it an experiment?
For each student, the study merely observed whether his or her school tested for drugs and whether he or she used drugs. So this was an observational study.
這是觀察研究還是實驗呢？
對每個學生而言，這個就只是觀察他／她的學校是否進行毒品測試；她／他是否使用毒品。這當然是個觀察研究。
An experiment would have assigned schools to use or not to use drug testing, rather than leaving the decision to the schools. For instance, the study could have "randomly" selected half the schools to do drug testing and half not to do it, and then a year later measured the student drug use in each school. As we will soon discuss, an experimental study gives the researcher more control over outside influences. This control allows more accuracy in studying the association.
進行實驗，會分配哪個學校進行毒品測試而哪個不會，而不是由學校自行決定。舉例來說，我們會隨機將學校分為兩半，一半進行測試，另一半則否。一年後才來對兩組不同學校的學生進行測試。實驗研究讓研究者可以更加地掌控外在的影響。這樣的掌控會讓研究的相關性更準確。
Advantage of Experiments over Observational Studies
實驗比觀察的好處
In an observational study, lurking variables can affect the results. The German study in Example 1 found an association between cell phone use and eye cancer. However, there could be lifestyle, genetic, or health differences between the subjects with eye cancer and those without it, and between those who use its own association with each of those cell phones a lot and those who do not. This could affect the association. For variables, example, a possible lifestyle lurking variable is computer use. Perhaps those who use cell phones a lot also use computers a lot. Perhaps high exposure to computer screens increases the chance of eye cancer. In that case, the higher prevalence of eye cancer for heavier users of cell phones could be due to their higher use of computers, not their higher use of cell phones.
觀察研究中，潛在的變數恨容易影響到結果。德國的研究顯示手機使用與眼癌有相關。其實，可能只是這些人的生活習慣、遺傳或健康有所不同而已。這都可能影響到相關性。舉例來說，電腦的使用可能就是一個潛在的變數。常使用電腦的人、同時也常手機，而暴露在電腦螢幕前也可能增加眼癌的可能性。所以，眼癌的高風險群可能是使用電腦而非使用手機導致的。
By contrast, an experiment reduces the potential for lurking variables to affect the results.
Why? We'll see that with a type of "random" selection to determine which subjects receive each treatment, the groups of subjects receiving the different treatments are "balanced": that is, the groups have similar distributions on other variables, such as lifestyle, genetic, or health characteristics.
實驗可以相對地減低這些變數的潛在可能性。怎麼說麼？可以從「隨機」選擇來決定哪個受試者被分配到哪個處遇。這些受試組別在受到各種處遇上是「平衡的」，意思就是說：在各種變數的分配上，如生活態度、遺傳或其他健康特質上，這些組別可以有相似的分配。
The two groups of mice have similar distributions on health.
這兩組老鼠在健康上就有相似的分配。
When the groups are balanced on a lurking variable, there is no association between the lurking variable and the explanatory variable.
當各組的在前在變數上是平衡的時候，在潛在變數跟依變相之間就不會有所相關。
Then, the lurking variable will not affect how that explanatory variable is associated with the response variable.
如此依來，潛在變數便不會影響自變項與一變項之間的關係。
Establishing cause and effect is central to science. But it's not possible to establish cause and effect definitively with observational studies.
建立因跟果是科學的中心。但是，依靠觀察研究是無法獲得決定性的因果關係的。
association does not imply causation.
相關性並非暗示著因果關係
Because it's easier to adjust for lurking variables in an experiment than in an observational study, we can study the effect of an explanatory variable on a response variable more accurately with an experiment than with an observational study.
因為，在實驗研究中比在觀察研究中比較容易去調整前在變數。我們在實驗中比在觀察驗就中更容易去判讀自變項對依變相影響。
If experiments are preferable, why ever conduct an observational study? Why bother to measure whether cell phone usage was greater for those with cancer than for those not having it? Why not instead conduct an experiment, such as the following: Pick half the students from your school "at random" and tell them to use a cell phone each day for the next 50 years. Tell the other half of the student body not to ever use cell phones. Fifty years from now, analyze whether cancer was more common for those who used cell phones.
如果實驗這麼完美，又何須觀察研究呢？又何必煩惱是否高手機使用的致癌率是否比低手機使用的致癌率大呢？又何必不做一個以下的實驗呢？隨機選擇學校中一半的人，叫他們每天使用手機50年；另一半則教他們千萬別用手機。50年後，再來研究使用手機者是否較容易得癌症？
Scientists often resort to observational studies. We'll study methods for designing a good observational study. This can yield useful information when an experiment is not practical.
科學家經常依靠著觀察研究。我們將會讀一些如何設計好的觀察研究的方法。這些方法可以在實驗無法實際上操作時產生一些好訊息。
Another reason observational studies are common is that many questions of interest do not involve trying to assess causality.
很多問題的重點不在於找到因果關係是另一個觀察研究如此普遍的原因。
Using Data Already available
使用已經存在的資料
The most readily available data come from your personal observations. Perhaps a friend recently diagnosed with brain cancer was a frequent user of cell phones. Is this strong evidence that frequent cell phone use increases the likelihood of getting brain cancer?
大部完成可獲得的資料來自於你的個人觀察。也許你最近一位診查出腦瘤的朋友是一個常使用手機的人。這算是一個強烈的證據來告訴我們說手機的使用會增加腦瘤的可能性嗎？
An association does not need to be perfect, however, to be causal. Not all people who smoke a pack of cigarettes each day will get lung cancer, but a much higher proportion of them will do so than people who are nonsmokers.
一個相關不一定要有完美的因果關係。不是每個每天抽一包菸的人都繪得肺癌，但相對於不抽菸的人來說，他們有比較高的可能性。
Instead of using anecdotal evidence to draw conclusions you should rely on data from reputable research studies.
避免使用個人的日常經驗來下結論，我們可以依賴那些聲譽良好的研究報告。
Results from well-designed studies are more Trustworthy than anecdotal evidence.
良好設計下所產生的結果比我們的日常經驗要可靠得多了。
The General Social Survey (GSS) is an example of a sample survey. It gathers information by interviewing a sample of subjects from the U.S. adult population to provide a snapshot of that population.
GSS就是一個抽樣調查的好例子。它利用對美國成人的抽樣調查，收集了能夠對這個母群體作一些反映的資訊。
A sample survey selects a sample of people from a population and interviews them to collect data.
抽樣調查從母群體中選擇了一些樣本，訪問他們並得到資料。
A sample survey is a type of observational study. The subjects provide data on the variables measured. There is no assignment of subjects to different treatments.
抽樣調查是一種觀察研究。受測者從變數的測量中提供資料。但，其中並無分配受測者不同的處遇。
Most countries conduct a regular census. This is a survey that attempts to count the number of people in the population and to measure certain characteristics about them. It is different from most surveys, which sample only a small part of the entire population.
大部分的國家都會進行普查。這個調查的主要目的是要算出母群體中有多少人，並對他們進行一些特質上的測量。這和一般只從母群體中抽少數樣本的的測量不同。
Although it's the intention of a census to sample everyone in a population, in practice this is not possible. Some people do not want to be counted. Some do not have known addresses for the Census Bureau to send a census form: Some people are homeless; others are transient.
雖然普查是想要對母群體中的每一個做抽樣，但是，在實務上這是不可能的。有些人不想被計算在內；有些人不在普查的地址中。有些狀況則是：有些人無家可歸，有些人只是短暫停留。
The U.S. Census Bureau and other government agencies continually take smaller samples of the population, rather than relying solely on the complete census.
美國普查局及一些政府機構持續地使用更小樣本的調查，而不願意使用完整的普查。
It is usually much more practical to take a sample rather than to try to measure everyone in a population. In fact, in the late 1990s statisticians at the Census Bureau proposed using sampling methods to estimate numbers of people missed by the ordinary census. Their proposal became a political battleground between Republicans (who did not want sampling used) and Democrats (who supported sampling).
在實務上，抽查一個樣本比去測量母群體中的所有人還要可行。事實上，在1990年代普查局的統計專家就建議使用抽樣方法來估計許多在通常的普查中會被遺漏掉的人。他們的提議變成了共和黨(不贊成抽樣)和民主黨(贊成抽樣)的政治戰場。
Post a Comment
<< Home