Judges, particularly appellate judges, do have significant power to make decisions on a partisan basis. This does happen. The United States Court of Appeals for the 9th Circuit (which includes California), for example, consistently rules more liberally than the United States Court of Appeals for the 5th Circuit (which includes Texas).
But, it is worth remembering that there are real limits to this tendency (and yes, I'm basically citing myself from dkospedia in this link, although the entry does have links backing this up):
[R]oughly 84% of [U.S.] Court of Appeal decisions (a sample year) are unanimous. This is consistent with a finding that 71% of U.S. Supreme Court decisions were unanimous from 1889 to 1959, that in a recent sample year 87% of California intermediate court of appeals decisions were unanimous, and that in the Missouri Supreme Court unanimous decisions are reached 98.5% of the time. Thus, while partisanship matters and influences results even in non-unanimous decisions, there is wide consensus on many legal issues.
About 91% of cases in the Colorado Court of Appeals are resolved with unpublished opinions or dismissals, which are generally unanimous. In addition, a large share of the 9% of Colorado Court of Appeals cases that are decided with published opinions are unanimous. Good statistics on the exact number of non-unanimouos published opinions are unavailable, but it is safe to say that the number probably less than a third of all cases. In other words, less than 3% of cases appealed to the Colorado Court of Appeals are decided on a non-unanimous basis.
Somewhere between 65%-75% of Colorado Supreme Court decisions are unanimous (it is hard to provide an exact percentage with certainty, as the only statistics that are easily available on the total number of dissenting opinions and some cases have more than one dissenting opinion), which is not too surprising because the state supreme court handles only the hardest cases.
Unlike the U.S. Supreme Court, which is fairly evenly divided on a partisan basis with a swing vote, however, the way cases resolved in the Colorado Supreme Court rarely unpredictable on a partisan basis, because the vast majority of dissents involve the same couple of judges (Coates and Eids, the more conservative judges on the Court, usually). You can count in the fingers of one hand number of non-unanimous cases in the Colorado Supreme Court each year, in which both Coates and Eids are in the majority.
Colorado Statistics cited in this post are found here.
The Last Word
It is also worth noting, given the natural tendency to focus on the U.S. Supreme Court and state supreme courts, just how powerful intermediate appellate courts are, while we are at it. The normal three judge panel that considers a federal appeal is almost always the last word on the matter (bracketed material and emphasis added below, plus one change in a percentage figure).
Appealling parties may ask the entire Court of Appeals in a circuit to second guess the decision of an individual panel (or in the 9th Circuit, a much larger panel of the court rather than the entire court). This is called an "en banc" review of a decision. Thus, a case can be appealled from a trial court to a three judge panel of the Court of Appeals, and then receive "en banc" review, and then go to the U.S. Supreme Court. [The en banc review step is optional.] En banc review, like U.S. Supreme Court review of a Court of Appeals decision, is discretionary and rarely granted. . . .
In 2002 in the 9th Circuit, for example, 801 published opinions were issued out of the thousands of cases that came before the court, of which about 40 were considered for en banc review, and 18 received en banc review (which in the 9th Circuit is actually a large panel of the court rather than the entire court). This was just 2% of cases heard on the merits.
En banc review is less frequent in absolute numbers, since the overall caseload is smaller, in other circuits. In 1999, a fairly typical year, only 94 cases in the entire federal system receive en banc review, which makes this level of review similar in frequency to, if not less common than, U.S. Supreme Court review. The vast majority of Court of Appeals cases decided on the merits (probably about 95% nationwide) are never even considered for en banc review.
Less than one case in a thousand filed in a U.S. Court of Appeals is reversed in en banc review or in review by the U.S. Supreme Court.
Colorado Appellate Courts often get the last word, as well. In the most recent fiscal year for which figures are available (2005-2006) there were 2622 appeals concluded in the Colorado Court of Appeals. In that same year 868 petitions in certioriari (i.e. appeals from Colorado Court of Appeals cases) where made to the Colorado Supreme Court, and 78 cases were decided on the merits with written opinions (a few of which would be in original proceedings direct from trial courts, such as water law cases).
A small number of cases may have been decided on the merits without a written opinion, and in a small number of cases, more than one petition for certiorari was resolved in a single opinion. Of course, some Colorado Supreme court cases affirm Court of Appeals rulings.
Thus, about a third of Colorado Court of Appeals cases are appealled to the Colorado Supreme Court, but the Colorado Supreme Court actually considers on the merits only about 3% of Colorado Court of Appeals cases, and probably reverses no more than 2% of those decisions.
The percentage of Colorado Supreme Court cases considered on the merits by the U.S. Supreme Court is on the order of 1%. Fewer than one in five thousand cases filed in the Colorado Court of Appeals will ultimately be reversed in the U.S. Supreme Court.